The Emergent Motherhood Mental Health Advantage: Did Pandemic Times Make a Difference? Open Access

Abstract

Research indicates that a new pattern of motherhood well-being advantage emerged in the 2010s for U.S. women. Although scholars have argued that maternal mental health worsened during the COVID-19 pandemic, whether the parenthood mental health gap changed during the pandemic is unclear. Using data from the National Health Interview Survey (N  =  29,241), this study examines the parenthood gap in yearly and quarterly changes in anxiety and depression during 2019–2021 for women aged 18–59, with attention to variation by partnership status. The results show that changes in anxiety and depression prevalence were similar across parental and partnership statuses, with indications that maternal advantages expanded among women who were single. In October–December 2020, anxiety prevalence increased more for single women without minor children of their own living in the household (“nonmothers”) than for single or partnered mothers. In April–June 2021, anxiety declined among mothers, especially single mothers, but remained higher than before the pandemic among single nonmothers. Some of these group differences in anxiety changes became nonsignificant after we controlled for household economic conditions, which were better in 2021 than in 2019 for all groups, particularly single mothers. In sum, trends in motherhood mental health advantages continued throughout the pandemic.

Introduction

Maternal stress and mental health are of great interest to demographers because these conditions have implications for child development, parenting quality, couple relationship quality, and fertility (Iacovou and Tavares 2011; Margolis and Myrskylä 2015; Tosi and Goises 2021; Turney 2012; Williams and Parra 2019). Earlier research tended to show that in the United States, mothers with minor children were more depressed than women without minor children (for a review, see Umberson et al. 2010), with some mixed evidence showing the importance of context, such as partnership status (Nomaguchi and Milkie 2003). However, recent research has shown that a clear pattern of a motherhood advantage in depression emerged during the 2010s (Nomaguchi and Milkie 2023).¹ Then, in the spring of 2020, the onset of the COVID-19 pandemic increased attention to the plight of mothers, whose responsibilities intensified as a result of the closures of most schools and daycare facilities (Adams et al. 2021; Augustine and Prickett 2022; Davis et al. 2021; Landivar et al. 2022; Petts et al. 2021). Numerous studies reported that mothers’ mental health worsened during the lockdown period from mid-March to May 2020 (see Table A1 for a list of selected studies; all tables designated with an “A” are in the online appendix),² and many claimed that mothers were disproportionally affected by various pandemic-induced stressors (Petts et al. 2021; Ruppanner et al. 2021; Zamarro and Prados 2021).

Did the motherhood mental health advantage disappear during the COVID-19 pandemic? Studies have not provided clear answers because of methodological limitations. First, most studies did not have an appropriate comparison group of women without minor children of their own living in the household (“nonmothers”). Findings from studies that included a comparison group of nonmothers are mixed (Table A1, highlighted in gray). Several studies showed worse mental health for parents than for nonparents (Elder and Green 2021; Fitzpatrick et al. 2020; Kowal et al. 2020; Montazer et al. 2022; Ruppanner et al. 2021; Zamarro and Prados 2021), a few studies reported no gap in mental health between those living with their own or any minor children and those who did not (Tull et al. 2020; Yan et al. 2022), and one study found that parents had lower psychological distress levels than adults without children (Lai et al. 2022).³ Second, almost all studies had no benchmark data collected before the pandemic, leaving it unclear whether mothers experienced a greater increase in anxiety and depression prevalence compared with nonmothers during the pandemic. In addition, most studies focused solely on the initial lockdown period, despite the continued changes to daily living induced by various public health measures, such as physical distancing six feet apart, mask-wearing, and restrictions on social gatherings (Moreland et al. 2020). The number of COVID-related deaths peaked in December 2020–January 2021 (Sabo and Johnson 2022). A turning point came in April–May 2021, when COVID-19 vaccines became available for all Americans aged 12 or older. However, many social distancing measures continued as variants of the virus (e.g., Delta, Omicron) spread across the nation from August 2021 through the rest of the year (Centers for Disease Control and Prevention 2024). Little research has tracked the parenthood gap in mental health changes throughout the pandemic. Third, because almost all studies used convenience samples or commercial online panels, the extent to which the findings can be generalized is unclear. A few studies using nationally representative samples showed that anxiety and depression increases among U.S. adults from 2019 to 2020 or 2021 were more modest than had been emphasized (Breslau et al. 2021; Jia et al. 2021; Kessler et al. 2022). These features of research on pandemic-related distress suggest the merit of assessing changes in the parenthood gap in mental health from 2019 to 2021 using nationally representative samples, which is a primary goal of the current study.

The findings of only modest increases in anxiety and depression prevalence among U.S. adults from 2019 to 2020 or 2021 might seem perplexing because of the widely held notions of heightened distress during that time. We consider three reasons why these findings push us toward a more careful consideration of the parenthood gap in mental health during the pandemic. First, most earlier studies focused on increases in stressors, ignoring positive changes that occurred during the pandemic (Milkie 2020; Prime et al. 2020). Increased resources during the pandemic, including the unusually generous expansions of the government social safety nets that started in April 2020 and extended throughout 2021 (Trisi 2023; U.S. Department of the Treasury n.d.), deserve more attention. Second, the emotional roller coaster experienced during the pandemic might not be captured by yearly observations. For example, Zamarro and Prados (2021) tracked biweekly changes during the lockdown period using nationally representative data and found that anxiety and depression prevalence among U.S. adults aged 18–65 increased sharply from March 2020 to early April 2020 but declined quickly by late May 2020. Examining biweekly trends from late August 2020 to late May 2021 using data from the U.S. Census Bureau Household Pulse Survey, Jia and colleagues (2021) found that anxiety and depression among U.S. adults aged 18 or older peaked during December 2020–January 2021, the period when the number of deaths from COVID peaked. These findings suggest the need to examine the relative increases and decreases in anxiety and depression prevalence with a shorter interval than yearly observations. Third, pandemic experiences were diverse among parents of minor children depending on other key factors (Torche and Nobles 2022; Vaterlaus et al. 2023). Given that some research found that the pandemic's effects on mental health were greater for those who did not live with a spouse or partner (Jace and Makridis 2021; Ray 2021; Thomeer 2022), the parenthood gap in mental health changes might depend on partnership status.

We use data from the National Health Interview Survey (NHIS) to accomplish three aims. Aim 1 is to assess whether the motherhood mental health advantage diminished during the pandemic. We examine the parenthood gap in the 2019–2021 yearly trends in anxiety and depression prevalence among U.S. women aged 18–59 before and after controlling for demographic characteristics and household economic conditions.⁴ In this study, we define “mothers” as women who self-report being parents (biological, adoptive, or stepparents) or guardians of minor children residing in their homes. We compare mothers with women aged 18–59 who do not have minor children of their own (including biological, adopted, or stepchildren) living in their households, whom we call “nonmothers” for simplicity. Aim 2 is to examine quarterly trends in 2019–2021, tracking relative increases and decreases in anxiety and depression prevalence during the pandemic years, to better capture short-term changes. Aim 3 is to investigate variation in the parenthood gap by partnership status, comparing the yearly and quarterly trends in mental health across four family types by parental and partnership statuses: partnered mothers, single mothers, partnered nonmothers, and single nonmothers.

We find that changes in anxiety and depression prevalence during the pandemic were modest and mostly similar for mothers and nonmothers, indicating that the maternal mental health advantage continued throughout the pandemic. The motherhood gap in mental health depends on partnership status: depression prevalence was lower for partnered mothers than for the three other groups throughout the three-year period around the pandemic. Unexpectedly, single mothers showed resilience to anxiety changes from 2019 to 2021, whereas single nonmothers experienced anxiety disadvantages during the pandemic. This study contributes to the literature on parenthood and well-being, the debates over subgroup differences in mental health difficulties during the COVID-19 pandemic, and the growing scholarship on singlehood.

Background

The Prepandemic Motherhood Gap in Mental Health: Theories and Trends

Rooted in role strain and role enhancement theories (Goode 1960; Sieber 1974), the demands–rewards perspective of the effects of parenthood on mental health posits that parenting experiences involve both challenges and rewards (Nomaguchi and Milkie 2020). The demands of caring for children involve physical, financial, emotional, and mental efforts and investments and can be major stressors for adults, especially when they perceive that their children's needs are more difficult to meet than expected (Deater-Deckard 2004). The rewards of caring for children involve aspects of care that can enhance adults’ social identity and social integration (Abrutyn and Mueller 2016). For example, young children reconnect adults to members of their family of origin (Aldrich et al. 2022; Min et al. 2022) and reorient adults’ social networks to local or virtual communities (Cossyleon and Geller 2022; Nomaguchi and Milkie 2003). The parenting role helps adults set priorities (Friedman et al. 1994) and encourages them to follow structured daily routines (Reczek et al. 2014). Most parents report that their children bring joyful and meaningful experiences to their lives (Nelson et al. 2013; Nomaguchi and Brown 2011). In short, parenting involves both burdens and rewards—a crucial perspective to consider in understanding parental status and mental health. Also important to consider is that not all individuals exposed to similar stressors develop poor mental health, partly because individuals can cope with challenging circumstances using financial, social, and psychological resources available to them (Pearlin 1999). Parenting demands might elevate anxiety and depression if parents lack resources, such as institutional supports for child-rearing, to cope with these demands (Aassve et al. 2015; Glass et al. 2016; Pollmann-Schult 2018; Stier and Kaplan 2020).

The life course perspective maintains that individuals’ lives are deeply embedded in the historical times in which they live (Carr 2018). Patterns of the parenthood gap in mental health can change, given that social changes might influence adults’ lives differently depending on whether they are parents of minor children or not. Previously, widespread notions contended that mothers with minor children were more depressed than nonmothers (Umberson et al. 2010). Yet, recent research found a new pattern of a motherhood depression advantage: mothers are, on average, less depressed than nonmothers (Metzger and Gracia 2023; Nomaguchi and Milkie 2023). For example, examining women aged 18–59 in the 1997–2018 NHIS, Nomaguchi and Milkie (2023) found that during the 2010s, depression prevalence increased for women who did not have children under age 18 living in their households but remained stable for women who did. The present analysis investigates whether the motherhood depression advantage disappeared during the COVID-19 pandemic while also examining another important mental health outcome: anxiety.

Demands and Rewards During the COVID-19 Pandemic: Variation by Parental Status

Much research on maternal mental health during the pandemic focused on the consequences of the sudden increases in childcare responsibilities as a result of school and daycare closures for work–family conflict during the lockdown period (McGoron et al. 2022; Montazar et al. 2022; Shockley et al. 2021; Yan et al. 2022; Zamarro and Prados 2021).⁵ Yet, other stressors are relevant, regardless of the presence of minor children: fear of contracting the virus (Elder and Green 2021; Jimenez et al. 2023), social isolation (Bierman and Schieman 2020), employment loss, financial strain, and food insecurity (Elder and Green 2021; Patrick et al. 2020; Rodrigues et al. 2021; Ruppanner et al. 2021). We discuss whether the degrees of increases in these stressors differed by parenthood status (i.e., the presence of one's own minor children in the home). We also consider, as the demands–rewards perspective encourages, whether differences in increases in rewards in these life domains may have varied by parenthood status.

Fear of contracting the virus was extremely high during the early stage of the pandemic, when information about how to prevent it was scarce. Misinformation was spread through social media, leading some people, for example, to sanitize grocery packaging in the hopes of preventing the transmission of the virus (Thomas and Feng 2021). Little research has examined the fear of contracting the virus as the major stressor influencing mental health.⁶ A few studies examined adults’ perceived threat of the COVID-19 pandemic regarding their daily lives and communities, their personal financial situation, and the U.S. economy. Elder and Green (2021) found that parents living with their children under age 18 were more likely than other adults to rate higher on the COVID-19 threat scale. Fitzpatrick and colleagues (2020) also found that adults living with children (of unspecified ages and relationship to the respondents) were more likely than other adults to rate higher on the COVID-19 threat scale. In contrast, Tull and colleagues (2020) found no difference in the perceived impact of COVID-19 between those who lived with children (again, of unspecified ages and relationship to the respondents) and those who did not. Two studies indicated that living with children was related to more exposure to the virus. Parents living with their minor children were more likely than other adults to report having had COVID (Elder and Green 2021). Lai and colleagues (2022) found that parents were more likely than those without children to know someone who had had the virus (although the child's age and whether residing with the parents were not specified). From these findings, we might expect that mothers’ mental health was more susceptible to the ups and downs of the pandemic than that of nonmothers. Therefore, we might expect to see greater increases in anxiety and depression prevalence among mothers than among nonmothers during the last quarter of 2020, when the number of COVID-related deaths was the highest.

Some researchers have argued that social isolation during the pandemic was particularly hard on caregivers of children (Brown et al. 2020; Lee et al. 2022). In contrast, other researchers have noted some positive consequences of social distancing orders on the daily lives of families with minor children (Milkie 2020; Prime et al. 2020; Vaterlaus et al. 2023). Drawing on in-depth interviews, Vaterlaus and colleagues (2023) highlighted variability in the effects of the COVID-19 pandemic on mother–child relationships, with roughly half of the mothers in their study reporting positive changes, approximately one fifth reporting a decline in mother–child closeness, and the rest reporting little change. Mothers in their study reported enjoying spending time with their children, especially when social distancing policies forced them to stay home and become more creative than usual in their time spent together. Marks and colleagues (2023), too, found that some parents viewed their relationships with their children during the lockdown period positively because they were working fewer hours for pay, allowing them to be more involved in children's lives than would have been the case otherwise. In their quantitative study, Lai and colleagues (2022) found that parents had more social interactions and were less stressed than adults who had no children during the lockdown period. Although we do not examine these constructs in the present study, these findings suggest that mothers were resilient during the pandemic and therefore were probably no more vulnerable to anxiety and depression than nonmothers during the pandemic.

Another major set of stressors was the economic effects of the pandemic, including layoffs, business shutdowns, temporary furloughs, and cutting work hours (Gassman-Pines et al. 2020; Zalewski et al. 2023). Some studies found that parents reported more financial concerns (Elder and Green 2021) or worry about losing their jobs (Ruppanner et al. 2021) than nonparents. Not often mentioned in these studies is that the U.S. government expanded social safety nets to an unusually large extent, beginning in April 2020 and continuing through the end of 2021 (U.S. Department of the Treasury n.d.). The American Rescue Plan involved numerous temporary policy changes, such as extensions in unemployment compensation benefits and the Supplemental Nutrition Assistance Program (SNAP), which reduced many Americans’ economic hardships (Center on Budget and Policy Priorities 2023). Notable was that most households with minor children received Economic Impact Payments (EIP) in March 2020, December 2020, and March 2021 (Cox et al. 2022). In addition, the Child Tax Credit (CTC) was expanded in terms of the maximum credit amount and coverages (Trisi 2023). Thanks to the EIP, the CTC expansion, and other program extensions, child poverty rates declined from 2019 to 2021 (Burns et al. 2022; Trisi 2023). Housing affordability increased in 2021 among mothers, especially among those with lower incomes (Pilkauskas et al. 2024). The percentage of adults who reported financial stability remained the same from 2019 to 2020 and slightly increased in 2021 (Board of Governors of the Federal Reserve System 2023). Although the EIP and CTC might have provided additional financial boosts to mothers relative to nonmothers, a direct comparison between mothers and nonmothers in terms of increases in financial strain and economic resources during the pandemic is difficult to make.

Variation by Parental and Partnership Statuses

Pandemic experiences varied across subpopulations of parents (Torche and Nobles 2022; Vaterlaus et al. 2023). Research shows that the anxiety and depression increases during the onset of the pandemic were greater among single than partnered individuals (Jace and Makridis 2021; Thomeer 2022), partly because living with significant others protected individuals from loneliness when social gatherings were restricted (Ray 2021). Having a spouse also mitigated the effects of job loss on increased anxiety and depression (Jace and Makridis 2021). Little research in the United States has considered how experiences of being single versus partnered differed during the pandemic depending on parental status, however. We examine variation across four family types by parental and partnership statuses: (1) partnered mothers, (2) single mothers, (3) partnered nonmothers (i.e., married or cohabiting women not residing with their own minor children), and (4) single nonmothers (i.e., not married or living with a partner and not residing with their own minor children). Research conducted before the pandemic found that single mothers experienced more parenting stress, fatigue, sadness, and depression than partnered mothers (Manuel et al. 2012; Meier et al. 2016; Nomaguchi et al. 2017). The wellness disadvantages of single mothers are largely attributable to greater challenges in household economic conditions for single mothers than partnered mothers (Harkness 2022) and could be reduced by institutional support (Pollmann-Schult 2018). In addition, fulfilling both parental and partnership roles might advantage partnered mothers in terms of their mental health because major social roles tend to provide a clear purpose in life through integrating individuals into the larger society, which is known to be a key factor in enhancing individuals’ mental health (Abrutyn and Mueller 2016). However, some research found that during the pandemic, partnered mothers experienced unequal increases in childcare and decreases in paid work activities compared with their male counterparts (Augustine and Prickett 2022; Petts et al. 2021). Although some evidence suggests that men increased domestic work more than women during the pandemic (Petts and Carlson 2024), mothers increased cognitive labor (i.e., anticipating, organizing, evaluating, and making decisions) related to domestic work more than fathers, which was negatively associated with mothers’ psychological well-being (Petts and Carlson 2023). Some research highlighted single mothers’ resilience during the pandemic (Taylor et al. 2022). In contrast, single women without children were most likely to experience social isolation among the four groups during the pandemic, but little research has investigated this group (Gao and Sai 2020). All in all, given the unusual increases in both stressors (e.g., childcare demands, fear of contracting the virus, social isolation) and resources (e.g., the expansions of the government economic assistance), it is unclear how these factors affected mental health differently for the four groups of women.

Potential Compounding Factors

The four groups of women by parental and partnership statuses vary considerably in demographic characteristics (e.g., age, race and ethnicity, immigration, and education), partly because patterns of childbearing and couple relationship stability vary by these characteristics (Smock and Schwartz 2020). In addition, the average hours of paid work and family income differ across the four family types (Passias et al. 2017). The degrees of changes in household composition and economic conditions (e.g., hours of paid work, family income, financial strain regarding health care, and food insecurity) during the pandemic might differ across the four family types. For example, young adults without children commonly moved back to their parental homes during the pandemic (Fry et al. 2020), which increased both conflict and support in the parent–child relationships (Gilligan et al. 2020). Anxiety and depression prevalence varies by these demographic characteristics and household economic conditions (Muntaner et al. 2013). Therefore, we examine the parenthood gap—and its variations across the four family types—in changes in anxiety and depression prevalence before and after the pandemic while controlling for demographic and household economic conditions.

The Present Study

The current study examines whether yearly (Aim 1) and quarterly (Aim 2) trends in anxiety and depression in 2019–2021 differed for mothers and nonmothers among U.S. women aged 18–59. On the basis of prior findings (Nomaguchi and Milkie 2023), we expect that anxiety and depression prevalence in 2019 (before the pandemic) was lower among mothers than among nonmothers (Hypothesis 1). Our literature review suggests that during the pandemic, some stressors increased more for mothers than for nonmothers, but other stressors increased less for mothers than for nonmothers. Relational rewards (which we cannot measure in the current study) and financial resources appeared to have increased more for mothers than for nonmothers. All things considered, we expect the degrees of the increases in anxiety and depression prevalence during the pandemic years and quarter years to be similar for mothers and nonmothers (Hypothesis 2). In addition, we examine variation in the parenthood gap by partnership status in both yearly and quarterly analyses (Aim 3). We expect that anxiety and depression prevalence in 2019 (before the pandemic) was lowest for partnered mothers among the four family types by parental and partnership statuses (Hypothesis 3) and that the increases in anxiety and depression prevalence during the pandemic were similar across the four family types (Hypothesis 4).

Methods

Data

We use data from the NHIS, a nationally representative, cross-sectional household interview survey conducted by the National Center for Health Statistics (NCHS; https://www.cdc.gov/nchs/nhis/). The NHIS collects data continuously from January to December each year, allowing us to assess quarterly trends in anxiety and depression prevalence. The NHIS randomly selects one sample adult aged 18 or older from each household. The sample sizes for the Sample Adult Interviews were 31,997 in 2019, 31,568 in 2020, and 29,482 in 2021.⁷ We selected respondents younger than 60 (n  =  19,590 in 2019; n  =  18,355 in 2020; and n  =  17,738 in 2021) who self-identified as women, resulting in the analytical sample size of N  =  29,241 (10,240 in 2019; 9,551 in 2020; and 9,450 in 2021). We imputed missing values by using a multiple imputation technique in SAS with 25 iterations (Berglund 2010; D'Agostino McGowan and Toll 2015).

Measures

Dependent Variables

Anxiety is a dichotomous variable assessed with the question, “How often do you feel worried, nervous or anxious? Would you say daily, weekly, monthly, a few times a year, or never?” Depression is a dichotomous variable based on the question, “How often do you feel depressed? Would you say daily, weekly, monthly, a few times a year, or never?” For both variables, those who reported daily or weekly are coded as 1; those who reported monthly, a few times per year, or never are coded as 0. These measures capture less severe anxiety and depression than clinically significant anxiety and depression.⁸

Independent Variables

Parental status is defined as being the parents or guardians of residential minor children or not. It is measured as a dichotomous variable created by using the three-category NHIS-composed variable. Those who are categorized in (a) “the sample adult is a parent of a child residing in the family” are coded as mothers ( = 1); those categorized in (b) “there are minor children residing in the family but sample adult is not their parent” or (c) “there are no minor children residing in the family” are coded as nonmothers ( = 0).⁹

Partnership status is a dichotomous variable. Those married or living with a partner are coded as 1 (the partnered), and others are coded as 0 (singles).

Family types include single nonmothers, partnered nonmothers, single mothers, and partnered mothers. They are measured with dichotomous variables of parenthood and partnership statuses.

Quarter year is measured with 12 dummy variables. For each of the three pandemic years, we measure the first quarter (Q1), January to March; the second quarter (Q2), April to June; the third quarter (Q3), July to September; and the fourth quarter (Q4), October to December.

Control Variables

Age is measured in years.¹⁰Race and ethnicity are measured by four categories: White, Black, Latina, and other race. Education ranges from 0  =  no education to 10  =  professional or doctoral degree.¹¹Foreign-born status is a dichotomous variable (1  =  born outside the United States, 0  =  born in the United States). Multifamily household is a dichotomous variable equal to 1 for a multifamily household and 0 for a single-family household. Household economic conditions include four indicators. Hours of employment in the previous week range from 0 to 95 hours (where 95  =  95+ hours). Family income–to-poverty ratio ranges from 0 to 10 (where 10  =  10+). Health care financial strain is the sum of seven items (α  =  .75) asking about the past 12 months (e.g., whether they had problems paying medical bills; 1  =  yes or 0  =  no); this variable ranges from 0 to 7. Food security is a 10-item composite variable indicating households’ food situation over the past 30 days (1  =  very low, 2  =  low, 3  =  marginal, 4  =  high; Coleman-Jensen et al. 2021).¹²

Sample Characteristics

Descriptive statistics for variables in the 2019–2021 pooled sample are shown in Table A3 for the total sample, by parental status, and for the four family types by parental and partnership statuses. Mothers and nonmothers differed in many demographic characteristics and household economic conditions. Mothers were, on average, younger than nonmothers.¹³ Mothers were more likely than nonmothers to be Latina and foreign-born. On average, relative to nonmothers, mothers worked fewer hours per week for pay, had a lower ratio of family income to poverty, and had less food security. Health care financial strain did not differ between the two groups. The four family types differed on most demographic characteristics and household economic conditions. On average, single mothers had the worst household economic conditions, followed by single nonmothers.

Table A4 presents means for indicators of household economic conditions in each year for the total sample, by parental status, and for the four family types. These data allow us to examine how the average household economic conditions changed from 2019 to 2020 and 2021 and whether the degrees of changes differed by parental status and across the four family types. Across the groups, the average weekly paid work hours decreased from 2019 to 2020 but returned to 2019 levels in 2021. The average family income–to-poverty ratio was higher in 2021 than 2019 for all subgroups except for single mothers. Across all four family types, the average score for health care financial strain was lower in 2021 than 2019, with greater changes for single mothers than single nonmothers. The average score for food security was higher in 2021 than 2019 for all family types except for partnered nonmothers, with greater changes among single mothers than among partnered mothers. Together, these statistics suggest that, on average, household economic conditions were better in 2021 than 2019 for both mothers and nonmothers and across the four family types, with some indications that single mothers experienced a bonus in food security relative to partnered mothers and partnered nonmothers and in health care–related financial situations relative to single nonmothers.

Analytical Plan

We conduct our analysis of yearly trends (Aim 1) in three steps. First, we chart anxiety and depression prevalence among mothers and nonmothers in 2019, 2020, and 2021 (Figure 1). To examine the parenthood gap in anxiety and depression prevalence in 2019, 2020, and 2021, we run two logistic regression models for each year for anxiety and depression, before and after controlling for demographic characteristics and household economic conditions (Table 1). Second, to assess whether changes in anxiety and depression prevalence from 2019 to 2020 or 2021 differed for mothers and nonmothers, we pool the three years of data and conduct a logistic regression model that includes year, parental status, and interaction terms between year and parental status (Table 2). We add demographic characteristics and household economic conditions to the model to examine whether the significant parenthood gaps in yearly anxiety and depression trends remain significant. Third, as part of Aim 3, we repeat the same sets of analyses for the four family types: partnered mothers, single mothers, partnered nonmothers, and single nonmothers (Figure 1, Table 2, and Table 3).

The analysis of quarterly trends (Aim 2) includes similar steps. First, we chart descriptive results for 2019–2021 quarterly trends in anxiety and depression prevalence among mothers and nonmothers (Figure 2). For each group, we conduct logistic regression models with and without controlling for demographic characteristics and household economic conditions for anxiety and depression (Table 4). Second, to examine whether differences in the quarterly trends by parental status are significant, we pool the three years of data and conduct a logistic regression model that includes quarter year, parental status, and interaction terms between quarter year and parental status (Table 5). We add demographic characteristics and household economic conditions to the models to investigate whether significant parenthood gaps in quarterly trends in anxiety and depression remain significant. The results using Q2 2019 as the reference group are presented.¹⁴ Third, as a part of Aim 3, we repeat the same sets of analyses for the four family types (Figure 3, Tables A6 and A7, and Table 6).

Following guidance from the NCHS (2020, 2021, 2022), we weight all analyses using the sample adult weights, which include design, ratio, nonresponse, and poststratification adjustments. To adjust the survey design, we use the PROC SURVEYMEANS command for descriptive statistics and the PROC SURVEYLOGISTIC command for multivariate analyses in SAS (D'Agostino McGowan and Toll 2015).

Results

Yearly Trends

Parenthood Gap

Descriptive results for 2019–2021 yearly trends in anxiety and depression prevalence among mothers and nonmothers are shown in Figure 1. As outlined earlier, statistical tests for the parenthood gap in anxiety and depression prevalence each year are conducted by logistic regression models presented in Table 1. The values labeled in Figure 1 are in bold when the parenthood gap in changes from 2019 to 2020 or 2021 is significant, for which the statistical tests are conducted by logistic regression models using the 2019–2021 pooled sample presented in Table 2.

Panel a of Figure 1 shows that in 2019, anxiety prevalence was lower for mothers than nonmothers (31.1% vs. 35.2%; also see Table 1, Model 1; b  =  −0.19, p < .001). In 2020, it was still lower for mothers than nonmothers (33.8 % vs. 37.6%; also see Table 1, Model 3; b  =  −0.17, p < .01), increasing from 2019 in similar ways for both groups; 2020 × mothers in Model 1 of Table 2 was nonsignificant (b  =  0.016, p > .10). In 2021, anxiety prevalence went back down for mothers (32.0%) but remained higher than in 2019 for nonmothers (39.3%), indicating an increase in the parenthood gap, which was marginally significant (see Table 2, Model 1; b  =  −0.121, p < .10 for 2021 × mothers, highlighted in gray). When we controlled for demographic characteristics and household economic conditions in the analyses conducted for each year (Table 1), the motherhood anxiety advantage became nonsignificant in 2019 and 2020 (Models 2 and 4)¹⁵ but remained significant in 2021 (Model 6; b  =  −0.15, p < .05). In the analyses using the pooled sample (Table 2), however, the parenthood gap in change in anxiety prevalence from 2019 to 2021 was no longer marginally significant after we controlled for household economic conditions (Model 2).¹⁶ Together, these results point to a small expansion of the motherhood anxiety advantage in 2021, but the degree of it was negligible when we held household economic conditions constant.

As shown in panel b of Figure 1, depression prevalence was lower for mothers than nonmothers in 2019 (9.0% vs. 12.7%; Table 1, Model 7; b  =  −0.39, p < .001), consistent with the motherhood depression advantages found in 2015–2018 (Nomaguchi and Milkie 2023). Notably, depression increased very little from 2019 to 2020 or 2021 for both groups,¹⁷ with no difference in the degree of changes between the two groups, given that the interactions (2020 × mothers and 2021 × mothers) in Model 3 in Table 2 were nonsignificant. When we controlled for demographic characteristics and household economic conditions, the motherhood depression advantages remained significant in 2019, 2020, and 2021 (Models 8, 10, and 12 in Table 1). Further, the degrees of changes in depression prevalence were similar for mothers and nonmothers (Table 2, Model 4).

In sum, depression prevalence was lower for mothers than nonmothers in 2019 and throughout the pandemic, suggesting that the motherhood depression advantages that emerged in the late 2010s (Nomaguchi and Milkie 2023) continued throughout the pandemic years (supporting Hypotheses 1 and 2). Anxiety prevalence was lower for mothers than nonmothers in 2019 at the descriptive level. Some indications point to an expansion of the motherhood anxiety advantages in 2021, but these differences became nonsignificant when we controlled for household economic conditions (supporting Hypothesis 1 and weakly supporting Hypothesis 2).

Variation Across Four Family Types

We find that the parenthood gaps in anxiety and depression depend on partnership status. Panel a of Figure 1 shows that in 2019, anxiety prevalence was lower among partnered mothers (29.8%) than among single mothers (35.3%), partnered nonmothers (32.8%), and single nonmothers (37.7%). (See Model 1 in Table 3 for significance tests.) After we controlled for household economic conditions, the anxiety advantages of partnered mothers relative to the three other groups remained significant (Table 3, Model 2) and continued in 2020 and 2021 (Table 3, Models 3–6). Examining changes from 2019 to 2021, Figure 1 (panel a) indicates that single mothers experienced a unique pattern: anxiety prevalence slightly decreased (although not significantly in terms of within-group comparisons). By contrast, it increased for nonmothers—notably for single nonmothers—and remained stable for partnered mothers. Single mothers’ unique pattern of decreased anxiety from 2019 to 2021 was significant compared with single nonmothers and marginally significant compared with partnered mothers (Table 2, Model 5; b  =  −0.291 for 2021 × single mothers, highlighted in gray). These differences became nonsignificant, however, when we held demographic characteristics and household economic conditions constant (Table 2, Model 6).¹⁸

For depression (Figure 1, panel b), in 2019, similar to anxiety, its prevalence was lowest for partnered mothers (7.0%), followed by partnered nonmothers (11.3%), whereas it was higher for single nonmothers (14.0%) and single mothers (15.3%). (See Model 7 in Table 3 for significance tests.) The depression advantages of partnered mothers compared with the three other groups remained when demographic characteristics and household economic conditions were held constant (Table 3, Model 8) and in 2020 and 2021 (Table 3, Models 9–12). Examining changes in 2021 showed that, like anxiety prevalence, depression prevalence became relatively higher for single nonmothers than for the three other groups, with the degree of the increases from 2019 significantly higher for this group than single mothers (Table 2, Model 7; b  =  −0.374, p < .05 for 2021 × single mothers, highlighted in gray). However, these differences became nonsignificant after controlling for household economic conditions (Table 2, Model 8).¹⁹

To summarize, anxiety and depression prevalence was lower for partnered mothers than for women in the three other groups throughout the three years (supporting Hypothesis 3), suggesting that the motherhood mental health advantages were concentrated among partnered mothers before and during the pandemic. In 2021, single mothers showed mental health resilience relative to the three other groups, whereas single nonmothers had anxiety disadvantages, especially compared with single mothers (contrary to Hypothesis 4). The differences between single mothers and single nonmothers in changes in anxiety prevalence remained significant after we controlled for household economic conditions.

Quarterly Trends

Parenthood Gap

We now examine quarterly trends in anxiety and depression prevalence for mothers and nonmothers. The descriptive trends are shown in Figure 2 (see Table A5 for the values for all quarters). Statistical tests for within-group quarterly changes are conducted by logistic regression models presented in Table 4, and those for the parenthood gap in quarterly changes are presented in Table 5.

For mothers, anxiety prevalence was relatively stable across the 12 quarters in 2019–2021, ranging from 28.7% to 35.6% (Figure 2; Table A5 for percentages; Table 4, Model 1, for statistical tests).²⁰ By contrast, for nonmothers, anxiety prevalence increased significantly from the latter half of 2020 to the end of 2021, peaking in Q4 2020 at 42.5% compared with 33.9% in Q2 2019 (Figure 2; Table A5; Table 4, Model 3).²¹ Model 2 in Table 4 indicates that anxiety prevalence for mothers would have been significantly higher in the latter half of 2020 and two quarters in 2021 than in Q2 2019 if household economic conditions did not improve in 2020 and 2021 compared with 2019, as we discussed earlier, most likely owing to the American Rescue Plan (see Table A4). For nonmothers, too, Model 4 in Table 4 suggests that anxiety prevalence would have been even higher if household economic conditions stayed the same throughout the six quarter years from Q3 2020 to Q4 2021.²² Nevertheless, these different patterns for mothers and nonmothers in anxiety trends across the pandemic quarters were significant only in Q2 2021, when anxiety declined more steeply for mothers than nonmothers (Table 5, Model 1; b  =  −0.380, p < .01 for Q2 2021 × mothers). The difference stayed significant when we controlled for demographic characteristics (Table 5, Model 2) but became nonsignificant when we also controlled for household economic conditions (Table 5, Model 3; b  =  −0.283, p > .10 for Q2 2021 × mothers).²³

Depression prevalence, like anxiety prevalence, was also quite stable across the pandemic quarter years for mothers, ranging from 7.4% to 10.4%; no quarter year saw significant increases (Figure 2; Table A5; Table 4, Model 5). For nonmothers, depression prevalence was significantly higher in Q4 2020 (17.4%) and Q2 2021 (15.1%) than in Q2 2019 (11.7%; Figure 2; Table 4, Model 7). When we controlled for demographic and household economic conditions, the quarterly trends in depression for mothers did not change (Table 4, Model 6); for nonmothers, depression prevalence in Q4 2021 (in addition to Q4 2020 and Q2 2021) became significantly higher than in Q2 2019 (Table 4, Model 8). However, these parenthood gaps in quarterly changes in depression were not statistically significant at the descriptive level (Table 5, Model 4) or after we controlled for household economic conditions (Table 5, Model 6).

In summary, the 2019–2021 quarterly trends in anxiety and depression were similar for mothers and nonmothers. During Q2 2021, anxiety decreased more for mothers than nonmothers, but the parenthood gap in these changes became nonsignificant after we controlled for household economic conditions—namely, perceived health care financial strain. These findings support Hypothesis 2.

Variation Across the Four Family Types

Finally, we examine variation across the four family types in quarterly trends in anxiety and depression. Descriptive trends are charted in Figure 3 (see Table A5 for the percentages for each quarter). Statistical tests for within-group quarterly trends are conducted in the logistic regression models presented in Tables A6 and A7. Those for between-group quarterly trends are conducted in models presented in Table 6.

Panel a of Figure 3 shows that the four family groups experienced increases in anxiety in different quarters between the last quarter of 2020 and the last quarter of 2021 (see Models 1, 3, 5, and 7 in Table A6 for statistical tests).²⁴ The interaction terms in Model 1 in Table 6 indicate that these group differences were significant in only two quarters, Q4 2020 and Q2 2021.²⁵ First, in Q4 2020, the increases in anxiety prevalence were significantly lower for partnered mothers than single nonmothers (Table 6, Model 1; b  =  −0.37, p < .05 for Q4 2020 × partnered mothers). After we controlled for household economic conditions (in addition to demographic characteristics), this interaction was no longer significant (Table 6, Model 3). Supplemental analyses suggest that the two groups did not differ in changes in household economic conditions, hence we could not identify which factor contributed to the smaller increase in anxiety for partnered mothers than single nonmothers during the period. Second, in Q2 2021, anxiety prevalence declined for partnered and single mothers (Table 6, Model 1; b  =  −0.38, p < .05 for Q2 2021 × partnered mothers; b  =  −0.73, p < .01 for Q2 2021 × single mothers) but remained high for single nonmothers. After we controlled for household economic conditions (in addition to demographic characteristics), these differences in changes in anxiety between partnered mothers and single nonmothers became nonsignificant in Q4 2020 and Q2 2021 (Table 6, Model 3),²⁶ whereas the differences in changes between single mothers and single nonmothers in Q2 2021 remained significant (Table 6, Model 3; b  =  −0.56, p < .05 for Q2 2021 × single mothers).

Panel b of Figure 3 shows that depression prevalence rose to higher than prepandemic levels for single nonmothers in Q4 2020 and Q2 2021 and for partnered nonmothers in Q4 2020. However, it did not increase for partnered or single mothers (also see Models 1, 3, 5, and 7 in Table A7).²⁷ Differences in the degree of changes in depression across the four groups in Q4 2020 and in Q2 2021 were not significant before or after we controlled for demographic characteristics and household economic conditions (see Table 6, Models 4–6, for interaction terms).²⁸

In sum, the 2019–2021 quarterly changes in anxiety and depression prevalence were mostly similar across the four groups, with small but notable differences observed in two quarters. In Q4 2020, the increases in anxiety were lower for partnered mothers than single nonmothers. In Q2 2021, anxiety prevalence decreased to prepandemic levels for mothers, especially single mothers, but remained higher than prepandemic levels for nonmothers, especially single nonmothers. These group differences in quarterly changes became nonsignificant when we controlled for household economic conditions, except for the steeper decline in anxiety among single mothers relative to single nonmothers in Q2 2021. These findings indicate anxiety disadvantages for single nonmothers and anxiety advantages for single mothers during the pandemic, contrary to Hypothesis 4.

Discussion and Conclusion

Using data from the NHIS, this study finds that contrary to the claim that mental health worsened more among mothers with minor children than among women without minor children of their own living in their households (referred to as “nonmothers” in this article) during the pandemic, mothers fared relatively well. The degrees of yearly (Aim 1) or quarterly (Aim 2) changes in anxiety and depression prevalence in 2019–2021 were mostly similar for mothers and nonmothers, with a small expansion of the parenthood gap in anxiety (favoring mothers) that became nonsignificant after we controlled for household economic conditions. We find that the parenthood gaps in mental health vary by partnership status (Aim 3). Across the three years, partnered mothers were less likely to experience anxiety and depression than the three other groups (i.e., single mothers, partnered nonmothers, single nonmothers). The analysis of quarterly trends reveals that the small expansion of the parenthood gap in anxiety was apparent during two quarters. One is the last quarter of 2020, when the United States had record numbers of COVID-19 deaths. At that time, anxiety worsened to a smaller degree for partnered mothers than for single nonmothers, although the group difference became nonsignificant after we controlled for household economic conditions. The other is the second quarter of 2021, when anxiety prevalence returned to its prepandemic level for mothers but not for nonmothers. Among the four groups, single mothers showed the greatest degree of changes, and single nonmothers showed the smallest degree of changes. The steeper decline in anxiety for single mothers than single nonmothers remained significant after we controlled for household economic conditions.

Our findings suggest that the trend of motherhood mental health advantages in the United States, which emerged in the 2010s (Nomaguchi and Milkie 2023), continued throughout the pandemic, supporting Hypotheses 1 and 2. Partnered mothers were less likely to experience anxiety and depression than women in the three other groups across the three years from 2019 to 2021, supporting Hypothesis 3. In terms of changes in anxiety and depression prevalence during the pandemic, single mothers showed greater resilience than single nonmothers in 2021, contrary to Hypothesis 4. Our analysis could not detect factors directly contributing to the unexpected resilience of single mothers relative to single nonmothers. By the second quarter of 2021, the third round of EIP had been distributed, COVID-19 vaccines were made available at no cost, and federal economic support programs (e.g., SNAP and CTC) were expanded (Center on Budget and Policy Priorities 2023; Cox et al. 2022). Our findings suggest the possibility that the development and universal access to COVID-19 vaccines at no cost and the increased institutional economic support helped reduce anxiety levels more for mothers than for nonmothers, with a greater gap among women who were single.²⁹ It is also possible that, as some studies indicated (Marks et al. 2023; Vaterlaus et al. 2023), mothers benefited from having more social interactions, including with their minor children, for everyday emotional support in the context of a pandemic relative to women without children present in their homes. Child-rearing responsibilities can lead adults to prioritize structured daily routines (Reczek et al. 2014), which could be especially beneficial during uncertain times. The finding of anxiety disadvantages among single nonmothers underscores the need for future research to investigate the well-being of single women living alone (Gao and Sai 2020).

The present analysis has limitations that future research should improve upon. It uses single questions to measure anxiety and depression, respectively, because they are the only questions asked in the 2020 and 2021 NHIS. As a common issue in studies using secondary analyses of public-use data from household-based surveys, the measure of “mothers” and “nonmothers” has limitations owing to the lack of information available on whether the sample adults have any children living outside the household, which does not allow researchers to distinguish these parents from childless adults. Still, household membership during the pandemic was especially relevant. The “nonmothers” category includes distinct groups whose experiences might differ—young women without children but who might have them in the future, mothers with minor children living elsewhere, older women who never had children, older mothers who have adult children in the household, and older mothers whose adult children live outside the home. As mentioned in the Methods section, we carefully considered these subgroups where possible and conducted sensitivity analyses. We urge researchers to consider who is included in the “nonmothers” category and distinguish it across these subgroups whenever feasible.³⁰

This study shows that increases in anxiety and depression prevalence among U.S. women aged 18–59 during the COVID-19 pandemic were more similar for mothers with minor children living in their homes and women without minor children of their own living in their households than prior studies suggest. The recent trends of maternal mental health advantages continued through the pandemic, with some variation by partnership status. The maternal mental health advantages in both anxiety and depression are generally concentrated among partnered mothers. During the pandemic, single mothers exhibited resilience against anxiety, whereas single nonmothers showed vulnerability. Future research assessing social factors contributing to the changing patterns in the parenthood well-being gap is crucial.

Acknowledgments

This research was supported by the Center for Family and Demographic Research, Bowling Green State University, which received core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P2CHD050959).

Notes

References