Financial strain, schooling modality, and mental health of U.S. adults living with children during the COVID-19 pandemic

Abstract

Background:

The COVID-19 pandemic upended contexts for families; relatively little work has studied the influence of rapidly changing contexts on the mental health of parents. We aimed to assess the relation between financial strain and schooling modality with the mental health of adults living with school-age children across the pandemic.

Methods:

Using a large, national sample from the COVID-19 Trends and Impact Surveys (CTIS; N=1,485,072 responses from November 2020 through June 2022), we used weighted multiple logistic regression with interactions for school semester to estimate changes in the association of frequent feelings of depression and anxiety, respectively, with financial strain and schooling modality, controlling for demographics, and state, across time.

Results:

In all time periods, financial strain was associated with reporting frequent feelings of depression and anxiety, respectively. The association grew over time (p<0.001) from aOR 2.25 (95%CI 2.19, 2.32)/aOR 2.63 (95%CI 2.54, 2.73) in Fall 2020 to aOR 3.11 (95%CI 3.01, 3.22)/aOR 3.79 (95%CI 3.64, 3.95) in Spring 2022. Living with children in fully online versus in-person schooling was associated with frequent feelings of anxiety and depression symptoms in all time periods, and increased from aOR 1.08 (1.05, 1.11)/aOR 1.06 (1.02, 1.10) in Fall 2020 to aOR 1.20 (1.10, 1.32)/aOR 1.28 (1.16, 1.42) in Spring 2022.

Conclusion:

Associations between financial strain and online-only schooling with poor mental health increased during the COVID-19 pandemic. Policies to support parents in the face of external stressors, such as economic instability and school closures, may improve overall population mental health.

Introduction

Over 12 million parents in the U.S. experience any mental illness in a given year (1), including depression and anxiety. Poor mental health of adults can affect the health of children,(2) creating a harmful cycle that affects multiple generations(3). Social stress theory suggests that exposure to stressors can weather an individual and lead to worse health over time.(6) Parenthood is a life stage associated with stressors, including high financial costs, which could help to explain high levels of depression and anxiety in early adulthood relative to older adulthood(7,8). The U.S. Department of Agriculture estimates that middle class families with a child born in 2015 will pay on average $233,610 per child on necessities like childcare, housing, food, and clothes.(9) Estimates including college tuition are even higher. It is no surprise, then, that parents report higher financial strain than non-parents.(10)

Parents faced additional stressors during the COVID-19 pandemic to provide foundational necessities for their families amidst economic variability and rapidly changing contexts of children’s schooling. In December 2020 and January – February 2021, adults with parent or caregiver roles reported higher levels of depression and anxiety than non-parents(1). Financial strain, which was heightened for families with lower income and jobs that were not able to be conducted remotely(11) contributed to cumulative stressors(12,13) that affected mental health of adults. In a national cross-sectional study among adults in Canada from May 14-29, 2020, Gadermann et al. found that parents’ self-reported mental distress was higher than the general population; stressors included financial strain, concern about family members catching the coronavirus, and challenges balancing work and childcare while schools and daycares were closed.(14) School closures also affected the economic circumstances of parents, who needed to stay home with children whose schools were closed.(15) Scholars called for the careful consideration of school closures on the mental health of children and their families,(16,17) but few studies have followed up to empirically document how factors like financial strain and school closures were associated with the mental health of parents or adults living with children in the household over the pandemic. Little work has explicitly explored the influence of financial strain and schooling modality on mental health and its evolution across time among parents or adults living with children during the whole pandemic period.

It is unclear how financial strain related to mental health among parents when controlling for demographics and schooling modality in the U.S. through June 2022 or how that relation changed over the course of the pandemic. The one study to our knowledge that approaches financial strain in parents while considering school closures in the U.S. was conducted by Moreland-Russell et al. Using pooled data from 3004 observations across 1913 participants from July 2020 through June 2021, they found that having children in hybrid schooling and that living in areas with lower median income were each associated with worse mental health for parents (18). However, they do not measure individual-level financial strain and end in June 2021. We extend the literature by using a large dataset with over 1 million responses across the U.S., extend the period of analysis through June 2022 to cover an extended pandemic period, and assess relations of individual-level financial strain on mental health across the pandemic. Informed by social stress theory and aiming to fill gaps in the literature by using a large dataset through June 2022, we aimed to explore changes in the relationship of financial strain and schooling modality with mental health of adults living with children 2020-2022 time period. To do so, we ask the following questions:

1. How were financial strain and schooling modality, two rapidly changing stressors during the COVID-19 pandemic, related to the mental health of adults living with children?

2. Did the associations between financial strain and schooling modality with adult mental health vary over time from 2020 through 2022?

Methods

Sample

We used data from a sample of U.S. adults surveyed through the COVID Trends and Impact Surveys (CTIS) between November 24, 2020, and June 15, 2022. The CTIS was a serial, cross-sectional survey led by the Carnegie Mellon University Delphi Research Group, designed in close collaboration with academic researchers and public health officials, and distributed via the Facebook platform. It collected over 20 million responses from April 6, 2020, to June 25, 2022. Respondents were asked about COVID-like symptoms, behaviors in the context of the pandemic, mental health, household context, and socioeconomic factors, with revisions made as new public health needs arose.

While the survey originally represents the population of active Facebook users in the U.S., we used survey weights provided by the survey administrators to align the sample with the U.S. population; the weights both adjusted for biases related to coverage and nonresponse and adjusted for demographic differences between Facebook users and the U.S. population. Further technical details on the survey instrument, sampling design, and weighting methodology can be found in Salomon et al.(19)

In this study, we were interested in the mental health of adults who lived with children under 18 in their household. To identify them in our sample, we included adults with a non-zero response to the question, “How many people under 18 years old are currently staying in your household?” We subset to responses collected during the school year (Fall 2020, Spring 2021, Fall 2021, and Spring 2022) from the inception of questions about in-person schooling (November 24, 2020). Responses for Fall 2020 were collected November 24, 2020 - December 19, 2020 (N=245,079). During this phase of the pandemic, vaccines were not yet available and while there was federal guidance from the CDC on safe re-openings for schools, state and local approaches varied widely. Responses for Spring 2021 were collected from January 3, 2021 - May 1, 2021 (N=995,497); during this period, vaccines were available for prioritized groups and then become available to all adults at different times in this period across states. Responses for Fall 2021 were collected August 29, 2021 - December 19, 2021 (N=49,631); responses for Spring 2022 were collected January 3, 2022 - May 15, 2022 (N=194,865).

Key covariates

Depression and anxiety

Frequent feelings of depression/anxiety were defined as binary outcomes, respectively, in response to: “In the past 5 days, how often have you…felt depressed / nervous, anxious, or on edge?” Responses were coded as positive if participants replied “more than half the time” or “all the time”, similar to other published work using these data(20). The questions were modified from the Center for Epidemiologic Studies Depression scale (CES-D)(21) and the General Anxiety Disorder (GAD) scale.(22) The CES-D is a validated scale used to assess the burden of depression symptoms at a population level; we use question 6 of the scale, which reads, “Please tell us how often you have felt this way during the past week…I felt depressed.” The GAD is a scale that has been validated against clinical diagnosis to capture symptoms of anxiety; we use question 1 of the scale, which reads, “Over the last 2 weeks, how often have you been bothered by the following problems? Feeling nervous, anxious or on edge.” The text of the CTIS question was updated from “past 7 days” to “past 5 days” starting March 2, 2021. We followed common practice to create a binary variable(20), to aid with interpretation and clinical relevance.

Financial strain and schooling modality

Financial worry was a binary variable defined by feeling “very worried” about their household's finances as opposed to “somewhat worried”, “not too worried”, or “not worried at all.” Schooling modality of children in the household was defined using a categorical variable with three options: in-person, online, or hybrid. School-related questions in the CTIS were introduced in Wave 5 of the instrument. From Waves 5 to 11 (November 24, 2020, to December 18, 2021) participants were asked, “Do any of the following apply to any children in your household (pre-K–grade 12)?” 1) Going to in-person classes full-time, 2) Going to in-person classes part-time”, with response options being “Yes”, “No”, and “I don’t know”. From Wave 12 until the end of the survey (December 19, 2021 to June 2022) participants were asked “Thinking about your oldest child under age 18, which of the following best describes their current schooling?”, with response options being: 1) “Going to in person classes”; 2) “Online, remote, or distance learning”, and 3) “Mix of in-person and online, remote, or distance learning”.

COVID-protective behaviors

We controlled for six COVID-protective behaviors, which were each defined as a response of “yes” to the following: 1) avoiding bars/restaurants/cafes in the past 24 hours, 2) avoiding large events in the past 24 hours, 3) avoiding market/grocery in past 24 hours, 4) avoiding work outside the home in the past 24 hours, 5) avoiding social contact, 6) wearing a mask in the past 24 hours.” We then created an index of them as a count variable (0-6) where 1=“yes” for each. These behaviors have been cited previously as being associated with poor mental health at the county level.(23)

Demographics

Individual-level demographic factors included 1) self-reported gender (Female, Male, Non-binary, Other/Prefer not to answer), 2) age group (18-24, 25-34, 35-44, 45-54, 55-64, 65+); 3) education level (Less than High school, High school, Some college, College/Professional degree, Graduate degree), 4) race and ethnicity (Asian, Black, Hispanic, White, Other/Multiple), and 5) state where respondent lived. Being unemployed was defined by a response of “no” to the question, "In the past 4 weeks, did you do any kind of work for pay?”

Analytic approach

First, we described the variables of interest by time period. Second, we estimated the weighted prevalence of frequent feelings of depression and of anxiety in each period. Third, we used individual-level survey-weighted logistic regression to model the probability of reporting frequent feelings of depression and of anxiety as a function of financial strain, schooling modality, demographics, COVID-protective behaviors, and state. We estimated a separate model for each of the four periods. Fourth, we tested for significant changes over time in these associations with time-interacted model specifications for the above regressions, where we used categorical variables for each time period. We include supplemental analyses using different model specifications (Poisson and linear regression, with numeric CED-D/GAD scores as outcomes) and removing COVID-19 protective behaviors as covariates.

Results

Table 1 describes the characteristics of the study sample across time. The study population’s demographics remained relatively stable over time, with slightly fewer people ages 18-24 responding in the final period than in the first. Financial strain and distribution of schooling modality fluctuated across the four time periods. Supplemental Table 1 shows the weighted prevalence of frequent feelings of depression and of anxiety by demographics, financial strain, and schooling across time. The weighted prevalence of financial strain (Supplemental Table 2) and schooling modality groups (Supplemental Table 3) by characteristics across time are shown in the Supplemental materials.

Table 1.

Characteristics of the sample by period

	Fall 2020	Spring 2021	Fall 2021	Spring 2022
Dates	November 24, 2020 - December 19, 2020	January 3, 2021 - May 1, 2021	August 29, 2021 - December 19, 2021	January 3, 2022 - May 15, 2022
Pandemic phase	No vaccines available	Early vaccine availability for priority groups	Widespread availability of vaccines	Widespread availability of vaccines
Sample size (complete cases)	245,079	995,497	49,631	194,865
Financial strain	56.9	49.3	43.3	43.8
Schooling modality
Full-time in-person	30.2	41.8	85.7	94.1
Hybrid	21.1	25.3	14.8	3.8
Online	56.0	42.5	10.3	2.2
Gender	(%)	(%)	(%)	(%)
Male	41.4	41.4	40.3	40.7
Female	56.0	55.6	55.9	56.0
Non-binary	0.7	0.8	1.1	0.9
Age group
Ages 18-24	9.8	9.2	8.0	1.7
Ages 25-34	19.9	18.3	19.9	18.8
Ages 35-44	34.4	35.4	36.3	42.8
Ages 45-54	24.5	25.1	25.0	28.4
Ages 55-64	7.5	7.8	7.3	6.5
Ages 65 plus	4.0	4.1	3.6	1.8
Race and ethnicity
Asian	2.9	3.4	3.0	3.2
Black	7.4	7.0	6.7	6.4
White	62.1	60.6	61.5	62.2
Hispanic	21.6	22.5	21.5	21.1
Multiple/Other	6.0	6.5	7.3	7.1
Education
Less than high school	5.2	5.4	5.1	4.5
High school	19.8	18.3	18.6	15.6
Some college	26.4	25.3	24.9	23.0
College/Professional degree	35.3	36.7	36.9	39.3
Graduate degree	13.4	14.3	14.5	17.6
Unemployed	31.9	32.9	29.2	27.0
Number of COVID-protective behaviors (mean, 0-6)	4.2	4.0	3.3	2.9

Note: CTIS data used: Fall 2020 collected November 24, 2020 - December 19, 2020 (N=245,079); Spring 2021 collected from January 3, 2021 - May 1, 2021 (N=995,497; Fall 2021 collected August 29, 2021 - December 19, 2021 (N=49,631)); Spring 2022 collected January 3, 2022 - May 15, 2022 (N=194,865). Number of COVID-protective behaviors (yes=1, no=0): 1) avoiding bar/restaurant/cafe in the past 24 hours, 2) avoiding large event in the past 24 hours, 3) avoiding the market/grocery in past 24 hours, 4) no work outside the home in the past 24 hours, 5) avoiding social contact, 6) wearing a mask in the past 24 hours

Table 2 shows the adjusted, survey-weighted odds ratios of frequent feelings of depression and anxiety by financial strain and schooling modality across time period. In Fall 2020, financial strain was associated with 2.25 (95%CI 2.19, 2.32) times the odds of anxiety and 2.63 (95%CI 2.54, 2.73) times the odds of depression. By Spring 2022, financial strain was associated with 3.11 (95%CI 3.01, 3.22) times the odds of anxiety and 3.79 (95%CI 3.64, 3.95) times the odds of depression. Having children in the home in online schooling relative to fully in person schooling was associated with 1.08 (95%CI 1.05, 1.11) times the odds of anxiety and 1.06 (95%CI 1.02, 1.10) times the odds of depression in Fall 2020. As of Spring 2022, living with children in online schooling was associated with 1.20 (95%CI 1.10, 1.32) times the odds of anxiety and 1.28 (95%CI 1.16, 1.42) times the odds of depression. Living with children in hybrid schooling was associated with poorer mental health in Spring 2022 [OR: 1.34 (95% CI 1.24, 1.44) for anxiety and 1.38 (95%CI 1.27, 1.5) for depression]. Findings are shown visually in Figure 1. Supplemental Table 4 presents full results, including coefficients for financial strain, schooling, COVID-protective behaviors, age, gender, race and ethnicity, education, and employment.

Table 2.

Adjusted odds ratios of frequent feelings of depression or of anxiety among adults living with children during the COVID-19 pandemic by financial strain and schooling modality across time.

	Fall 2020		Spring 2021
	Anxiety	Depression	Anxiety	Depression
Financial strain	2.25 (2.19, 2.32)***	2.63 (2.54, 2.73)***	2.37 (2.34, 2.41)***	2.67 (2.63, 2.72)***
Hybrid schooling	0.99 (0.95, 1.03)	0.98 (0.93, 1.03)	1.03 (1.00, 1.05)*	1.03 (1.00, 1.05)*
Online schooling	1.08 (1.05, 1.11)***	1.06 (1.02, 1.1)**	1.15 (1.14, 1.17)***	1.07 (1.05, 1.09)***
In person schooling	ref	ref	ref	ref
	Fall 2021		Spring 2022
	Anxiety	Depression	Anxiety	Depression
Financial strain	2.99 (2.81, 3.19)***	3.69 (3.42, 3.99)***	3.11 (3.01, 3.22)***	3.79 (3.64, 3.95)***
Hybrid schooling	1.00 (0.86, 1.16)	1.09 (0.91, 1.29)	1.34 (1.24, 1.44)***	1.38 (1.27, 1.5)***
Online schooling	1.18 (1.08, 1.3)***	1.16 (1.04, 1.3)**	1.2 (1.1, 1.32)***	1.28 (1.16, 1.42)***
In person schooling	ref	ref	ref	ref

Notes: CTIS data used. Complete case analysis used. All models control for financial strain, schooling modality, age, gender, education, race and ethnicity, employment, state, and number of COVID-19 avoidant behaviors. Number of COVID-protective behaviors (yes=1, no=0): 1) avoiding bar/restaurant/café in the past 24 hours, 2) avoiding large event in the past 24 hours, 3) avoiding the market/grocery in past 24 hours, 4) no work outside the home in the past 24 hours, 5) avoiding social contact, 6) wearing a mask in the past 24 hours. Referent categories: Not reporting financial strain; being employed in the last four weeks; and fully in-person schooling. p<0.05 *, p<0.01 **, p<0.001***

Figure 1. Adjusted odds ratio for frequent feelings of anxiety and of depression by financial strain and schooling modality.

Note: CTIS data used: Fall 2020 collected November 24, 2020 - December 19, 2020 (N=245,079); Spring 2021 collected from January 3, 2021 - May 1, 2021 (N=995,497; Fall 2021 collected August 29, 2021 - December 19, 2021 (N=49,631)); Spring 2022 collected January 3, 2022 - May 15, 2022 (N=194,865). All models control for age, gender, education, employment, race and ethnicity, state, and COVID-protective behaviors.

Table 3 shows the odds ratios from models with a categorical time period interaction coefficient for each of our covariates of interest, in order to test for changes over time in the strength of their association with worse mental health. Interaction terms showed that the association between worse mental health and financial strain strengthened in each of Fall 2021 and Spring 2022, respectively from Fall 2020 (p<0.001). The association between hybrid schooling and worse mental health strengthened in Spring 2022 relative to Fall 2020 (p<0.001). The association between online schooling and worse mental health strengthened in Spring 2021 (p<0.001) for anxiety and in Spring 2022 (p<0.01) for anxiety and (p<0.05) for depression relative to Fall 2020.

Table 3.

Time interaction for the adjusted odds of frequent feelings of depression or of anxiety across financial strain and schooling modality across the pandemic.

	aOR (CI, 95%)
	Anxiety	Depression
Time Period 2	0.66 (0.59, 0.74)***	0.76 (0.67, 0.87)***
Time Period 3	0.76 (0.6, 0.98)*	0.88 (0.67, 1.17)
Time Period 4	1.16 (0.98, 1.36)	1.09 (0.91, 1.31)
Financial worry	2.25 (2.19, 2.31)***	2.63 (2.54, 2.73)***
Financial worry * Time Period 2	1.05 (1.02, 1.09)**	1.02 (0.98, 1.06)
Financial worry * Time Period 3	1.33 (1.24, 1.42)***	1.4 (1.29, 1.53)***
Financial worry * Time Period 4	1.38 (1.32, 1.44)***	1.44 (1.37, 1.52)***
Unemployed	1.02 (0.99, 1.04)	1.22 (1.18, 1.26)***
Unemployed *Time Period 2	1.06 (1.02, 1.09)***	1.03 (0.99, 1.07)
Unemployed *Time Period 3	0.98 (0.91, 1.06)	0.94 (0.86, 1.03)
Unemployed *Time Period 4	1.07 (1.02, 1.12)**	0.99 (0.94, 1.05)
Online schooling	1.09 (1.06, 1.12)***	1.07 (1.03, 1.11)***
Online schooling * Time Period 2	1.06 (1.03, 1.10)***	1.0 (0.97, 1.04)
Online schooling * Time Period 3	1.1 (1.0, 1.22)	1.1 (0.98, 1.23)
Online schooling * Time Period 4	1.11 (1.01, 1.22)*	1.19 (1.07, 1.33)***
Hybrid schooling	1.01 (0.97, 1.05)	0.98 (0.94, 1.03)
Hybrid schooling * Time Period 2	1.02 (0.98, 1.07)	1.05 (1, 1.11)
Hybrid schooling * Time Period 3	0.99 (0.84, 1.16)	1.1 (0.91, 1.32)
Hybrid schooling * Time Period 4	1.33 (1.22, 1.45)***	1.41 (1.28, 1.55)***

Note: Data pooled from all 4 periods. All models control for age, gender, education, race and ethnicity, and state. Responses for Fall 2020 (Time Period 1) were collected November 24, 2020 - December 19, 2020 (N=245,079); responses for Spring 2021 (Time Period 2) were collected from January 3, 2021 - May 1, 2021 (N=995,497); responses for Fall 2021 (Time Period 3) were collected August 29, 2021 - December 19, 2021 (N=49,631); responses for Spring 2022 (Time Period 4) were collected January 3, 2022 - May 15, 2022 (N=194,865). Time was coded as a 1-4 categorical variable and an interaction term was added for each covariate of interest in the model, and also for demographic variables (i.e., gender, race and ethnicity, education level, age group, state). Number of COVID-protective behaviors (yes=1, no=0): 1) avoiding bar/restaurant/cafe in the past 24 hours, 2) avoiding large event in the past 24 hours, 3) avoiding the market/grocery in past 24 hours, 4) no work outside the home in the past 24 hours, 5) avoiding social contact, 6) wearing a mask in the past 24 hours. p<0.05 *, p<0.01 **, p<0.001***

Supplemental analyses (Supplemental Figures 2 and 3; Supplemental Table 3) produced similar findings.

Discussion

This paper shows that amidst rapidly changing contexts and stressors during the COVID-19 pandemic, financial strain and schooling modality were independently and increasingly associated with poorer mental health among adults living with children, even when controlling for sociodemographic characteristics. First, financial strain was consistently associated with frequent feelings of depression and anxiety across the entire study period. Second, hybrid or online schooling were associated with frequent feelings of depression and anxiety relative to in person schooling full time, when controlling for financial strain and demographic factors. Third, the strength of association between financial strain and online schooling with worse mental health each grew significantly over the course of the pandemic. These findings build on a literature assessing parental mental health and well-being (24) bearing evidence to the high and increasing toll of financial strain on mental health during the COVID-19 pandemic.

These findings are consistent with social stress theory, whereby exposure to cumulative stressors(6) contribute to worsened mental health. Findings from this study build on past literature on financial stain and mental health(25) more generally, expanding this work through 2022 using a large national sample of parents and adults living with children in the household, a group that faced unique and multiple changing contexts and stressors during the COVID-19 pandemic. This work is also consistent with a growing body of literature that documents the importance of multiple assets in protecting health(26). Access to assets changed rapidly during the pandemic, including stability in routines and having a safe place for children to go outside the home. Having more assets (through for example income or through safe care and physical infrastructure for children during the day) may have protected mental health for parents. In their study of adults through February 2021, Moreland-Russell et al. found that parents with lower median area-level income were more likely to report worse mental health during the pandemic.(18) They also found that parents whose children were in hybrid learning had worse mental health than parents whose children were in in-person schooling.(18) However, that study ended in June 2021 and only included employed parents (n=3004). Our study extended one year farther through June 2022, when schooling policies and routine engagements changed dramatically due to access to vaccines and improved understanding of the coronavirus; additionally, our study included 1,437,586 responses over the total period and included unemployed parents. In another study of parents during the pandemic, in a sample of hourly service employees in New York with children ages 2-7 years, Gassman-Pines found that reporting two or more hardships during the pandemic (such as job loss, income loss, caregiver burden, or household member feeling sick) was associated with negative parental mood.(27) Financial strain was heightened for families with lower income and jobs that were not able to be conducted remotely(11) and parents may have had altered labor opportunities (15) as a result of needing to stay home with children during school closures. These findings, including our own, contribute strong evidence to the role of cumulative stressors, including financial strain and not having stable in-person schooling for children, in shaping parental mental health during the pandemic with potential spillover effects beyond the pandemic period.

Federal and state policies intended to support families may have prevented even greater financial strain and poor mental health in the U.S. For instance, the expanded child tax credit provided additional support to parents during the pandemic but reports on its effects on parental mental health have been mixed, potentially due to the temporary nature of the tax benefits. Using the serial cross-sectional Household Pulse Survey, Batra et al. found lower reporting of symptoms of anxiety and depression among low income parents who received the child tax credit;(28) also using the Household Pulse Survey, Nam and Kwon reported a reduction in anxiety symptoms among low-income parents receiving the child tax credit(29); however, Glasner et al. found no short term effect of the child tax credit on symptoms of anxiety or depression in adults with children when using the longitudinal Understanding America Study.(30) In this context, our finding that the association between economic context and mental health increased over time suggests, too, that evolving economic contexts, including high inflation and expiration of child tax credit support (which ended in December 2021), may have been important factors in parental mental health over the pandemic.

There are six main limitations to consider. First, this study is a large, serial cross-sectional study that did not link participant responses over time; as such we are not able to assess longitudinal trends within individuals. However, the intent of the study was to assess how the associations between financial train and schooling modality with depression or anxiety changed over time. Future work could more formally address causal pathways of the observed associations with parental mental health; longitudinal data on individuals would be particularly useful for those studies. Second, the CTIS questionnaire include a question about whether participants were parents or legal guardians of someone under 18 years only starting in December 2021 and did not include the relation between adults and the children with whom they lived for the rest of the study period; thus, it is possible that some members of our sample were not parents to the children in their household. However, the size and breadth of our data provide a unique opportunity to study the mental health of all adults living with children in the household during the pandemic, plausibly most of them being parents. Third, the CTIS included brief screening tools for depression and anxiety based off of full-scale screening instruments, the CES-D(21) and the GAD(22) which have been validated against clinical samples and used in other published works(20); formal diagnoses should come from a provider. However, the mental health questions employed allow for broad scale examination of mental health across a large sample during a time period of great need for some assessment of population mental health. Fourth, it is possible that the relation between adult mental health and child schooling went in both directions; towards the end of the study period in Spring 2022, it is possible that parents had the choice of selecting schooling mode and that parents who were more anxious or depressed were more likely to select online or hybrid instead of in-person schooling. It is possible that the direction of causation may have changed over the course of the pandemic. Fifth, it is possible that Facebook users may not reflect the general adult U.S. population in ways not captured by the variables included in the weights that aim to increase sample representativeness (31). However, the focus of this paper was to estimate the associations between financial strain and schooling modality with mental health, not their national prevalences, and it is unclear whether or how the relationship between these two factors and mental health would systematically differ between Facebook users and non-Facebook users, which mitigates this concern. Sixth, it is possible that other factors, such as the ability to work remotely, may have influenced mental health of parents; however, our inclusion of financial strain and education in our models may help mitigate some of this bias, given associations of remote work and higher education, associated with lower financial strain.

Conclusion

Our main findings highlight the importance of financial strain as a determinant of mental health and the importance, from a policy perspective of, consistently, and not only temporarily, supporting parents and families, given their particular stressors. Decisions and policies around schooling modality–or of how to best support families and communities during periods of school closure–may consider parental mental health and economic context as additional factors. Presence of disparities across parental well-being prior to the pandemic suggest that while these forces existed before the global pandemic, they were likely intensified after it. Moreover, social stress theory suggests that stressors will remain beyond the global pandemic period. In sum, given the importance of adult mental health on children’s mental health (as well as a host of other life outcomes), addressing the economic and social contexts of adults living with children can help protect population mental health.

What is already known on this topic –

While others have documented changes in the mental health of children in different periods of the COVID-19 pandemic, few studies have explored the role of rapidly changing contexts on the evolution of mental health of adults living with children during an extended pandemic period.

What this study adds –

Using a large national sample with over 1 million responses, financial strain and online-only schooling were associated with frequent feelings of anxiety and depression among adults living with children during the COVID-19 pandemic—and these associations grew stronger over time.

How this study might affect research, practice or policy –

These findings suggest that the economic context of families should remain a policy priority and that the implementation of school closures should be accompanied with support mechanisms for families in future emergencies.