their children were at risk of contracting and spreading COVID if they returned to school. Fall 2020 produced the highest wave of COVID deaths of the entire pandemic in the United States (Worldometer n.d.), and although some parents had a strong desire to return children to in-person school, fears about COVID in schools remained high, and many people were hesitant to potentially expose their children to others (Horowitz and Igielnik 2020).

Whether children's school modalities in fall 2020 matched parents' concerns about COVID is unclear. There is substantial reason to think they did not. On the one hand, the opportunity to send children to school in person remained somewhat limited, particularly in school districts with a high percentage of racial-ethnic minority children (Landivar et al. 2022). Therefore, a good number of parents who wanted to send their children to school may not have had the option to do so. On the other hand, the limited number of remote-eligible jobs and the paucity of paid family leave and other work-family care supports in the United States likely led some parents who may have preferred to keep children home to send them to in-person school anyway because they lacked other viable options (Calarco 2024). Indeed, women's labor force participation was directly related to children's in-person school attendance (Landivar et al., 2022).

Using data on 709 partnered parents of schoolage children (age 5-17) from the Study on Parents' Divisions of Labor during , the aim of this study is to examine how children's school modalities in fall 2020 are associated with parents' perceived stress and how this may have varied by parents' concerns about COVID-19. It is likely that some parents' stress was reduced by the return to in-person school, particularly among those who were less concerned about COVID. Among parents who were concerned about COVID, keeping their children in virtual school may have mitigated stress, whereas others may have experienced heightened stress if they had to send their children to school in person despite their concern. The lack of overall change in parents' stress levels in fall 2020, therefore, is likely the result of high degrees of incompatibility between school modality and concerns about COVID.

Stress rapidly increased among the general U.S. population at the start of the COVID-19 pandemic (Twenge et al. 2021). U.S. parents in particular struggled to cope with the consequences of social lockdowns and reported greater risks of increased stress and psychological distress (Lamar et al. 2021;McGoron et al. 2021;Twenge et al. 2021). Specifically, prior to the pandemic, only 3.5% of U.S. parents reported high parenting stress, but by spring 2020, that percentage had jumped to 22.4% (Adams et al. 2021). A study from June 2020 found more than a quarter of U.S. parents reported worsening mental health (including increased stress) since the outbreak in March 2020 (Patrick et al. 2020;Zamarro and Prados 2021), with mothers exhibiting the worst declines (Calvano et al. 2022;Cheng et al. 2021;Johnson et al. 2021). Stress among parents during the early days of the pandemic has also been linked to other issues, such as increases in parental burnout and alcohol use (Calvano et al. 2022;Kerr et al. 2021;Lamar et al. 2021) and higher incidence of adverse childhood outcomes and stress among children (Calvano et al. 2022;Cohodes, McCauley, and Gee 2021).

Greater stress among parents at the beginning of the pandemic may stem from several sources, including increases in domestic responsibilities, social isolation, and financial strain. To reduce the risk of transmitting the novel coronavirus at the beginning of the COVID-19 pandemic, U.S. officials closed all schools, forcing every school-age child to receive their education at home (Van Lancker and Parolin 2020). Losing in-person school was noted by parents as a major factor in the stress and burnout that they experienced early in the pandemic (Calvano et al. 2022). Indeed, the amount of time parents dedicated to homeschooling their children was positively associated with the level of stress parents reported (Cuadrado et al. 2021;DesRoches et al. 2021). Conflict between partners also increased for homeschooling parents, especially when one parent-more often the motherreported doing the majority of the homeschooling (Basso et al. 2023). Working parents and single parents were especially stressed given their increased responsibilities at home (Carlson, Petts, and Pepin 2022;Mullins, Scutelnicu, and Charbonneau 2020;Rehman, Smith, and Poobalan 2021). Yet despite the link between homeschooling and parents' stress early in the pandemic, U.S. parents' reports of stress and psychological distress in fall 2020-when many schools returned to in-person instructionremained significantly higher than they were prepandemic and not substantively different compared to the start of the pandemic in spring 2020 (Adams et al. 2021;Carlson, Fielding-Singh, et al. 2022;Twenge et al. 2021).

In addition to the stress of homeschooling children, parents' stress likely also increased due to social distancing. To reduce the spread of COVID-19, the CDC encouraged people to socially isolate themselves and to quarantine when they began experiencing symptoms (American Red Cross 2020). Pre-COVID research shows that quarantine and social isolation are related to many mental health issues, such as increased risk of posttraumatic stress disorder (Sprang and Silman 2013). For some, these negative consequences may last years (Brooks et al. 2020). Social isolation and quarantine during the COVID-19 pandemic was no different. Thirtythree perecent of adults reported high levels of distress (e.g., anxiety, sleeplessness, depression, and loneliness) when social distancing protocols did not end after the expected two weeks in March 2020 (Keeter 2020a(Keeter , 2020b)). One study found depression levels to be higher among parents who had to be quarantined for two weeks than among those who did not and especially so for parents of young children and those of lower socioeconomic status (Wu et al. 2020). Social distancing impacted not only individuals but also entire family systems at the start of the pandemic given that parents reported feeling angry and frustrated with their children more often while isolating (Johnson et al. 2021).

Lastly, the financial strain from reducing work hours to fulfill their increasing childcare responsibilities likely also contributed to stress among parents, especially mothers (Cheng et al. 2021). Indeed, compared to other women, mothers, especially those of school-age children, were much more likely to reduce paid work hours or leave the workforce altogether early in the pandemic (Heggeness et al. 2021;Petts, Carlson, and Pepin 2021). Working mothers who were primarily responsible for homeschooling children were 3 times more likely to exit the labor force (Petts et al. 2021), and their male partners were more likely to report an increase in alcohol use during the pandemic (DesRoches et al. 2021), suggesting that fathers too were negatively impacted by the loss of childcare supports and the strains this produced.

Persistently high levels of stress among parents in fall 2020 likely stemmed not only from the state of the pandemic but also from disconnects between parents' desires for particular school modes (i.e., in person/virtual) and their ability to access and/or utilize them. Notably, parents' preferences about school modality were largely shaped by their concern about COVID. Decisions about schooling in fall 2020 took place at the same time as the second wave of COVID and the highest daily death rates of the entire pandemic in the United States. (Worldometer n.d.). Heightened stress during this time (Twenge et al. 2021) was likely shaped by this spike in cases, with parents worrying about exposure, fears about dying from COVID after contracting it, caring for ill family members with COVID, experiencing the hospitalizations and deaths of friends and loved ones, and dealing with the disruptions to life caused by positive tests and need for quarantine. In short, although school reopening coincided with increased politicization of the pandemic, many people in the United States remained highly concerned about COVID in fall 2020 (Carlson, Fielding-Singh, et al. 2022), and research shows these worries were strongly tied to stress (Carlson, Fielding-Singh, et al. 2022).

Parents in particular had to worry about how their children's school modalities could expose their families to COVID. Therefore, in considering the relationship between children's school modes and parents' stress in fall 2020, COVID concerns stand out as a potentially important moderating factor. Indeed, compared to mask mandates, social distancing requirements, or access to childcare, U.S. parents' concern about COVID health and safety was the strongest predictor of preferred school format (Limbers 2021). Parents concerned about COVID generally preferred virtual options, whereas parents with no or minimal COVID concerns generally preferred in-person schooling (Moreland-Russell et al. 2022). At the beginning of June 2020, approximately one-third of U.S. parents planned to keep their children in distance learning, and around half said they would definitely be sending their children to in-person school in the fall (Kroshus et al. 2020). Later estimates suggest that among parents who had the option to send their child to school in person, only two-thirds chose to do so (Henderson et al. 2021).

Although parents' COVID worries predicted their preferred school mode, parents' options were constrained not only by public policy-25% of school districts remained remote-only in fall 2020-but also family, work, and care circumstances. There is evidence of significant overlap between COVID worry, exposure risk, school reopening policies, and parental school mode preferences such that districts with populations highly concerned and at risk for COVID (e.g., urban, politically liberal, Black, Hispanic) were more likely to begin the fall 2020 school year remote-only (Chua et al. 2021;Landivar et al. 2022). However, there is also evidence that disconnects between parents' COVID worries and their children's actual school mode may have been fairly common. More than half of parents with children attending in-person school in fall 2020 reported feeling concerned about their children being exposed to COVID-19 at school, with 20% of parents reporting being very concerned (Horowitz and Igielnik 2020).

According to stress process theory, social conditions are not inherently stressful. Rather, a condition's stress potential depends not only on how one views it but also on one's estimation of their ability to effectively handle the condition (Lazarus and Folkman 1984;Pearlin 1989). The hallmark of stressors is that they are appraised by the individual as unwanted, undesirable, and/or unexpected (Lazarus and Folkman 1984). As such, incompatibilities between parents' preferred and actual school modalities for their children may have been a stressor. That is, parents likely appraised children's school modes as stressors-and experienced elevated stress levels-both when parents were worried about COVID but children attended school in person or when parents were not worried about COVID but children were learning remotely.

Incompatibilities between COVID concerns and school mode may have also produced stress via parents' role conflicts. Role theory (Barnett 2008;Coverman 1989) states that role conflicts occur when the demands of roles across multiple statuses are incompatible, leading to stress. Whether the result of public policy or driven by other factors, parents' inability to match their children's school mode with their levels of COVID worry likely produced conflicts between parent and work roles.

Parents concerned about COVID who sent their children to in-person school may have had conflicts between work and parenting obligations, leaving them little choice but to send children back to school either because they could not afford to stay home with them or because they could not find suitable childcare. Although remote work options expanded drastically during the pandemic, essential frontline industries, which are not remote eligible, constitute one-third of all private sector jobs (U.S. Bureau of Labor Statistics 2021). Additionally, the percentage of workers working from home dropped between spring and fall of 2020 from 69% to 57% (Saad and Wigert 2021). This was driven by a substantial drop in the percentage of workers working from home exclusively (from 54% to 35%). Although remote work can contribute to role conflict (e.g., Yucel and Chung 2023), it does provide concerned parents an option to keep their children in virtual school and reduce virus exposure, which may ultimately reduce stress (Carlson, Fielding-Singh, et al. 2022).

Lack of access to paid leave likely also inhibited working parents from choosing virtual schooling. Although the Families First Coronavirus Response Act provided employees working for eligible employers with up to 10 weeks of paid leave if children's schools were closed, those with in-person school options did not qualify. And even if one did qualify, length constraints meant that leave would not cover even one semester of the school year. Moreover, the lack of paid family leave for most workers in the United States (U.S. Bureau of Labor Statistics 2023) meant that parents would have to quit their job, take unpaid time off (something few could afford), or use workplace vacation/sick leave policies (if available) to stay home with children unless someone was already ill with COVID.

In addition to work-family conflict, some parents may have felt conflicted in their parenting role, struggling with concerns for their children's safety on one hand and concerns for their children's development on the other. For some parents, the detriments of remote learning may have outweighed the risk of COVID-19, leading them to send children back to school in person, even if reluctantly. Estimates indicate that three in five parents felt that the first year of the pandemic had a negative effect on their child's education and emotional well-being (Braga and Parker 2022), and many were adamant about their child returning to school, citing health and academic development as key concerns (Anderson, Hughes, and Trivedi 2021). Indeed, evidence from the end of 2020 shows a profound decrease in children's educational performance and learning following COVID lockdowns (Duckworth et al. 2021;Kuhfeld et al. 2020).

Among parents with low COVID concerns who did not have the option to send their children to inperson school, similar role conflicts were also possible. Because many of the parents in this group likely had virtual education imposed on them by living in school districts that remained fully remote, role conflicts likely occurred as parents had to continue managing children's homeschooling as they did in spring 2020. Moreover, the disconnect between parents' preferences and options likely also produced stress because parents had little control over their situation and thus may have been more likely to struggle coping with this stressor.

Although much of the conversation about school modalities in fall 2020 is about in-person school compared to virtual schooling, approximately 20% of school districts opened in a hybrid format (i.e., part-day, rotating, part-week, and other schedules; Landivar et al. 2021). How hybrid learning affected parents' stress and how this may have varied by COVID concerns is not clear, however. On the whole, hybrid formats do not have clear benefits over other learning formats. Although they may have helped reduce parents' concerns about learning loss, resulting in lower stress compared to a fully virtual format, they also increased the risks of exposure to COVID. Additionally, although hybrid formats alleviated some of the responsibilities of educating children from parents, they did not necessarily resolve work-family conflict because partday and part-week in-person learning may have been incompatible with parents' full-time work schedules. Moreover, hybrid and in-person school schedules were variable and changing as outbreaks led schools to revert to full remote learning at times (Parks et al. 2021). As such, hybrid schooling may have created schedule instability for parents. Schedule instability is linked to parenting stress generally (Craig and Brown 2017;Luhr, Schneider, and Harknett 2022), and evidence shows that COVID-related disruptions to schooling and work specifically had substantial impacts on parent stress (Calarco 2024). Overall, the benefits of hybrid models may have been offset by their drawbacks, indicating no discernable stress differences of hybrid models from other modes of learning. As to whether the association between hybrid formats and parents' stress varied by parents' COVID concerns, this is also unclear. Yet given that hybrid formats did not fully serve either those highly concerned about COVID or those with low COVID concerns, stress among parents with children in hybrid formats and stress differences compared to other school modes may not have varied substantially by parents' COVID concerns.

The aim of this study is to examine how children's school modalities in fall 2020 are associated with parents' perceived stress and how this varied by parents' concerns about COVID-19. Our conceptual framework and review of the literature leads to the following hypotheses:

Hypothesis 1a: Parents with children attending school in person full-time in fall 2020 reported less stress than parents with children attending school virtually. Hypothesis 1b: Stress levels for parents with children in a hybrid school format in fall 2020 did not differ significantly from parents with children attending in person full-time or virtually. Hypothesis 2: Parents who were more concerned about COVID-19 in fall 2020 reported higher stress than parents who were less concerned. Hypothesis 3: Incompatibility between children's school modalities and parents' COVID concerns are associated with higher levels of stress. Specifically, stress should be highest among the following groups of parents: Hypothesis 3a: children mandated to learn virtually; low concerns about COVID Hypothesis 3b: children learning in person; high COVID concerns. Stress should be lowest among these groups of parents: Hypothesis 3c: children learning in person; low COVID concerns; Hypothesis 3d: children learning remotely by choice; low COVID concerns. DATA AND METHODS Data Data for this study came from the Study of U.S. Parents' Divisions of Labor during COVID-19 (SPDLC). The SPDLC (Carlson and Petts 2022) is an opt-in online survey assessing changes in families during the pandemic that has both a longitudinal panel and repeated cross-sectional design. At each wave of the study, previous respondents are resurveyed, and new respondents are recruited. The sample is comprised of married or cohabiting U.S. parents with at least one biological child in the home (at the time they enter the study). Respondents provide information for themselves and their household. Information on parenting behaviors and children is gathered for a single focal child-the youngest child in the household. Wave 1 of the survey was administered in April 2020 and included 1,157 participants. Wave 2 of the SPDLC was administered in November 2020. A total of 828 respondents (68.6%) from Wave 1 also completed Wave 2. An additional 1,138 new respondents participated in Wave 2 of the survey. In all, 1,966 unique respondents completed the Wave 2 survey. Analytic Sample This study was limited to parents of school-age focal children (age 5-17) who participated in the Wave 2 survey. Of the 1,966 returning and new participants who completed the Wave 2 survey, 746 reported having their youngest child attending K-12 school either in person, hybrid, or remote for the 2020-2021 school year. Listwise deletion of cases with missing information on study variables (n = 28) and exclusion of cases where parents report their or their partner's gender as something other than male or female (n = 9) resulted in a final analytic sample of 709 parents.

Dependent variable. The dependent variable in this study is a measure of perceived stress, derived using Cohen's Perceived Stress Scale (Cohen, Kamarck, and Mermelstein 1983). Perceived stress was a summary scale (range = 0-40; α = .91) of 10 questions with responses ranging on a scale from 0 (never) to 4 (very often).

Independent variable. The focal child's school modality in fall 2020 was the primary independent variable. Participants were asked: "In what format is your child currently attending school?"-in person full-time, in person part-time and virtual/e-learning part-time, and virtual/e-learning full-time. Respondents were also asked if this school option was the only one available to their child or one of many options. Because all school districts in the United States provided families with a full virtual learning option in fall 2020 but not every district provided an in-person learning option, we combined information from these two survey questions and assessed children's school mode using a four-category variable: mandated virtual (reference), virtual by choice, hybrid, and in person.

was a summary scale (range = 0-12; α = .76) of three Likert-style questions ranging from 0 (strongly disagree) to 4 (strongly agree). The questions included: "I worry that someone I know will contract the coronavirus," "I am afraid to go out in public due to the coronavirus," and "The coronavirus outbreak is being blown out of proportion." Responses were coded such that higher scores indicated greater concern. Confirmatory factor analysis indicated that each standardized factor loading was .67 or higher, which along with the Cronbach's alpha of .76, indicated acceptable scale reliability. To more simply and clearly demonstrate the distribution of COVID concerns, we also constructed an ordinal variable for COVID concern with three categories-no COVID concerns (scale score of 3 or less), ambivalent about COVID (score of 4-8), and concerned about COVID (score of 9 or higher).

Control variables. We included several control variables that potentially affected stress, school modality, and COVID concerns in multivariate analyses. Respondents' and their partners' gender were dichotomous variables (1 = female). Education level of the respondent and partner were categorical variables with indicators for "high school or less" (reference group), "some college," "associate's degree," "bachelor's degree," and "advanced degree." Race-ethnicity of respondent and partner were operationalized as a series of dummy variables for "White" (reference group), "Black," "Hispanic," and "Other." Respondents' political orientation was categorized as "conservative" (reference group), "liberal," and "moderate." Respondents' religion included the following categories: "Protestant" (reference group), "Catholic," "agnostic," "none/atheist," and "other." Household income was a seven-item ordinal scale assessing total monthly household income after taxes.

Models also included controls to account for access to schedule and workplace flexibility for both the respondent and their partner. Schedule flexibility included the ability to select which days of the week or daily hours employees worked, and workplace flexibility included employees' ability to work from home. Because not all parents were employed or could work flexibly/remotely, measures of job flexibility were measured with categories for "out of labor force" (reference), "working-no access to schedule/workplace flexibility," "access to schedule flexibility only," "access to work from home only," and "access to both schedule flexibility and work from home."

Other control variables included living with extended family members (1 = yes), the number of children under 18 living in the household, living with a person with a chronic illness (1 = yes), and age of youngest child (ages 5-12 years = 0; ages 13-17 = 1). We dichotomized child age because numerous analyses indicate that parents of younger school age children faced greater difficulties than those of teenage children (e.g., American Psychological Association 2020; Collins et al. 2021).

Lastly, the study included controls for mental health diagnosis and whether anyone in the household had a confirmed case of COVID-19 (1 = yes). To construct these variables, we took data from both Wave 1 and Wave 2 of the SPDLC. Upon first interview, all participants were asked "Have you ever been diagnosed with a mental illness" and "Has anyone in your household had a confirmed case of COVID?" Follow-up respondents were subsequently asked whether this had changed since the last survey. To create variables for mental illness diagnosis and household COVID cases, participants who answered "yes" in either Wave 1 or Wave 2 on each item were coded as "yes" (yes = 1) for each variable.

Data analysis was conducted using Poisson regression. We ran two sets of regressions, one without statistical controls and one with statistical controls. Poisson regression was chosen for our analyses because the distribution of stress scores was positively skewed and all values of perceived stress were positive integers. Because school modality and concern over COVID-19 may be confounded with each other as predictors of parents' stress, regression analyses were run in four steps. Model 1, which tests Hypotheses 1a and 1b, included school mode alone. Model 2, which tests Hypothesis 2, included concern over COVID-19 alone, and Model 3 included both variables. To test Hypotheses 3a through 3d, Model 4 included an interaction term for school mode and concern about COVID-19. Models 5 through 8 repeated the aforemtnioned steps but included the full set of control variables.

Table 1 presents descriptive statistics for all variables. On a scale of 0 to 40, the mean reported stress level of parents was 16.06. The mean value for concern about COVID was 7.43 out of 12.00. Just under one-half of parents (45%) reported being concerned about COVID-19 (scale score of 9 or higher), and 15% reported no concerns (scale score of 3 or lower). Twenty-seven percent of parents reported that their youngest child was attending school in person in fall 2020, with about 20% attending in a hybrid format. Of the 53% of parents who reported their youngest child was attending school virtually full-time at the time of the survey, a slight majority reported choosing this option.

Table 2 presents results from Poisson regression analyses for our main variables of interest. For full results for models including controls, see Appendix Table A1 in the online version of the article. Consistent with Hypothesis 1a, results from Model 1 show significant bivariate differences in parents' stress levels by children's school modalities. Parents with children attending school in person report significantly less stress than those with children in mandated virtual school and those who chose virtual learning (who had the most stress, on average). Results in Model 1 provide mixed support for Hypothesis 1b. Parents with children in a hybrid format had stress levels similar to those with children attending in person and lower than those who chose virtual learning. Differences compared to parents with children in mandated virtual education, however, were not significant at conventional levels (p < .10).

Results in Model 2 provide support for Hypothesis 2, showing that greater concern about COVID-19 is associated with higher levels of stress. When both school modality and concern about COVID are included in the same model (Model 3), no differences in parents' stress are observed across school modality, suggesting that general differences in stress by school modality are due to variations in COVID concern. Indeed, parents who chose virtual education reported the greatest COVID concerns, followed, in order, by parents with children in mandated virtual, hybrid, and inperson school modes (see Table 3).

Nevertheless, estimates for statistical interaction between school mode and COVID concern in Model 4 provide substantial evidence for Hypothesis 3-that school mode and concern about COVID-19 interact in predicting parents' stress. The significant interaction coefficients for virtual by choice (b = .034; p < .001), hybrid (b = .030; p < .01), and in person (b = .039; p < .001) indicate that stress differences between those whose children were mandated to learn virtually and other parents vary significantly at different levels of COVID concern. Moreover, although no association between school mode and stress is found after introducing controls (Model 5) and the association of COVID concern with stress is reduced by nearly half (Models 6 and 7), the statistical interaction between school mode and COVID concern in predicting parents' stress is little changed and remarkably robust to the inclusion of statistical controls (Model 8).

To better understand how COVID concerns moderate the association between school modality and parents' stress, predicted marginal mean values from Model 8 of Table 2 (holding all controls at their mean values) are graphed in Figure 1. Table Table 1. Descriptive Statistics (N = 709). Variable M/% SD Variable M/% SD Perceived stress 16.06 7.52 Participant religion School modality Protestant 25.81 Mandated virtual 24.26 Catholic 23.41 Virtual by choice 28.91 Agnostic 13.54 Hybrid 20.31 None/atheist 22.85 In person 26.52 Other 14.39 Concern about COVID-19 7.43 3.39 Household income Concerned 44.99 Up to $2,999/month 17.35 Ambivalent 40.20 $3,000-$6,999/month 43.44 No concern 14.81 $7,000+/month 39.21 respondent gender respondent job flexibility Female 56.14 Out of labor force 22.85 Partner gender No flexibility 15.51 Female 44.85 WFH access only 10.72 respondent education Schedule flexibility only 9.31 HS or less 7.76 Both WFH and flex 41.61 Some college 16.50 Partner job flexibility Associate's degree 10.72 Out of labor force 21.86 Bachelor's degree 38.50 No flexibility 29.90 Advanced degree 26.52 WFH access only 12.98 Partner education Schedule flexibility only 11.42 HS or less 16.93 Both WFH and flex 23.84 Some college Associate's degree 18.48 10.30 N umber of children younger than age 18 1.64 .73 Bachelor's degree Advanced degree 31.88 22.43 L iving with someone with a chronic illness 25.53 respondent race-ethnicity White 76.87 r espondent political orientation Black 6.77 Conservative 33.43 Other 8.89 Moderate 24.54 Latino/a 7.48 Liberal 42.03 Partner race-ethnicity Living with extended family 6.77 White 76.30 Age of youngest child Black 6.21 5-12 years old 72.01 Other 9.44 13+ years old 27.99 Latino/a 8.04 Diagnosed with mental illness 17.92 C onfirmed case of COVID in household 3.39 Source: Data come from Survey on Parents' Divisions of Labor during COVID-19. Note: HS = high school; WFH = work from home. A2 in the online version of the article shows comparsions of predicted marginal means within and across school modality. The results illustrated in Figure 1 support Hypotheses 3a through 3d. Among parents who were largely unconcerned about COVID, those with children in mandated virtual school had the highest levels of stress (Hypothesis 3a), and those with children attending in person (Hypothesis 3c) or virtually by choice (Hypothesis 3d) exhibited the lowest stress levels. By comparison, at the highest levels of concern, parents who had children attending school in person report the highest levels of stress (Hypothesis 3b), which was significantly higher than parents with children in mandated virtual school but not virtual by choice. Among those most concerned about COVID, parents with children in hybrid school report significantly higher levels of stress than parents with children in mandated virtual school but no differences compared to parents with children in person or those who chose virtual schooling.

To better understand the consequences of school reopening policies for parents' stress in fall 2020, Table 3 shows the distribution of COVID concern across school modality. For simplicity, we present results using a categorical measure of COVID concern (no COVID concern, ambivalent about COVID, and concerned about COVID).

As Table 3 shows, incompatibilities between parents' COVID concerns and their children's school modalities, which are associated with greater stress, were more frequent when children were attending in-person school than when mandated to learn virtually. Thirty-seven percent of parents with children attending in person reported being concerned about COVID versus 13% of parents with children mandated to virtual schooling who reported no COVID concerns. Relatedly, we find that compatibility between COVID concerns and school modes-a circumstance that should be optimal to parents' mental health-was less frequent when children attended in-person school. Only 23% of parents with children learning in person reported no COVID concerns. Comparatively, 44% of parents whose children were mandated to learn virtually had concerns about COVID. Based on parents' COVID concerns, then, in-person education in fall 2020 was less likely than mandated virtual education to benefit parents' stress and more likely to exacerbate it. Indeed, parents whose children were mandated to be online were 238%, or 3.38 times, more likely to experience compatibility between their COVID concerns and their children's school mode than incompatibility (44% vs. 13%), whereas parents with children attending in person were 35% less likely to experience compatibility between their concerns and their children's school modes as they were to experience incompatibility (23% vs. 37%).

The COVID-19 pandemic increased parents' stress (Cheng et al. 2021;Johnson et al. 2021;Lamar et al. 2021;McGoron et al. 2021). Although many factors contributed to this, the move to remote learning in spring 2020 was cited as a primary factor (Calvano et al. 2022;Cuadrado et al. 2021;DesRoches et al. 2021), leading some officials to advocate heavily for school reopening in fall 2020 as a means of reducing parents' burdens. Overall stress levels, however, did not rebound substantially when schools reconvened (Adams et al. 2021;Calvano et al., 2022;Carlson, Fielding-Singh, et al. 2022;Twenge et al. 2021) despite 55% of students having access to full-time in-person school in fall 2020 (Landivar et al. 2022). This study examined the role of school modality in parents' stress in fall 2020 by considering how parents' levels of concern about COVID-19-a known stressor during the pandemic (Carlson, Fielding-Singh, et al. 2022)-may have moderated the association between school mode and parents' stress.

Analyses showed that parents' level of concern about COVID-19 was positively associated with stress. Bivariate results showed that parents with children learning in person exhibited less stress than parents with children learning virtually. Yet after accounting for COVID concern, which was significantly related to parents' stress and school mode, no significant differences in parents' stress levels were found across school modality. Consistent with stress process theory, results from interaction tests revealed that the association of school modality with parent stress varied by parents' concerns about COVID-19 such that incompatibilities between COVID concerns and school mode frequently resulted in higher levels of stress among parents. For parents with low COVID concerns, in-person school was associated with the lowest levels of stress, and mandated virtual school was associated with the most stress. Conversely, among those with high COVID concerns, mandated virtual school was associated with the lowest levels of stress, and in-person school was associated with the highest levels of stress.

Although sending children to school in person appears to have reduced stress among a limited number of parents who had no concerns about COVID, the benefits of in-person schooling for stress appear to be outweighed by the costs given that more parents with children learning in person reported being concerned about COVID than not. The opposite was true of mandated virtual schooling. Although a small proportion of parents unconcerned with COVID appeared stressed by being forced to educate children virtually, 3 times as many concerned parents appear to have experienced less stress by having children in mandated virtual learning compared to in-person learning. Overall then, these findings demonstrate that parents' stress likely did not decrease substantially in the United States in fall 2020 despite increased access to in-person school because in-person education was more frequently incompatible with parents' COVID concerns than mandated virtual schooling.

Although mandated virtual education was associated with significantly less stress for concerned parents than in-person school, the same was not true for parents who chose virtual education compared to in-person school. Data limitations do not allow us to fully examine reasons for this finding, yet a few explanations seem probable. First, parents may have chosen virtual learning for their children over in-person learning even if they faced serious work-family conflicts due to the presence of other stressors, such as an immunocompromised family member, that may have tempered the benefits of virtual learning for stress. Such a conclusion is supported by the fact that parents who chose virtual learning reported the most concern about COVID in our sample and, as supplemental analyses revealed (not shown), were also the most likely of all parents to have household members with a chronic illness. Second, compared to those with virtual learning mandates, parents with in-person school options may have received less public, social, and workplace support for continuing to educate children at home, leading to higher levels of stress. Additionally, parents with options who kept children home may have agonized over their choice or received pressure from children to return to in-person education. Future research should explore the sources of stress among those who chose to keep children home once in-person school options became available.

Our analyses revealed that COVID concerns were significantly tied to school mode-with concerned parents more likely to choose virtual schooling. Yet many highly concerned parents sent children to school in person anyway. Reasons for this could include subpar virtual learning options, lack of remote work and paid leave options, and financial strain from mothers working fewer hours while managing their child's at-home schooling in spring 2020 (Carlson, Fielding-Singh, et al. 2022;Chua et al. 2021). Many parents struggled to manage the tasks of working from home and schooling their children virtually at the start of the pandemic (Cheng et al. 2021), which could have led concerned parents to still "elect" to send their children to inperson school despite virtual schooling options being available to them (Landivar et al. 2022). Understanding exactly why parents concerned about COVID sent children to in-person school is another important topic for future research.

For parents with high levels of concern about COVID who had little choice but to send children back to school, in-person school appears to have exacerbated their stress. Although every district offered a virtual option in fall 2020, the paucity of necessary work-family supports that would enable parents to choose this option-paid family leave, flexible schedules and work from home options, and greater virtual learning support from schoolsmeans parents did not or could not take advantage of this option. These supports in conjunction with virtual learning could have instead lowered parents' stress levels by allowing parents to school their children effectively in the home while reducing the risk of infection.

Although this study helps us understand why school reopening was not associated with appreciable overall decreases in parents' stress in fall 2020, there are some limitations to consider. First, the study design of the SPDLC may limit generalizability of the findings. The SPDLC is an online opt-in survey of partnered U.S. parents where information was gathered on a single focal child's schooling circumstances. Although estimates from multivariate regression analyses using opt-in panels are comparable to those produced from random samples (Levay, Freese, and Druckman 2016), the focus on partnered parents means the results do not necessarily speak to the experiences of nonpartnered parents for whom stress may have been greatest in fall 2020 (Rehman et al. Poobalan 2021). Additionally, because the SPDLC gathered data on the youngest child only, parents with both preschool and schoolage children are not included in these analyses, and information about other children's school modalities are absent from the data. Whether this aspect of the study design affects results is unclear. Although the loss of childcare certainly affected the mental health of parents of preschool aged children, research suggests that pandemic strains were most acute among parents of school-age children, especially younger school-age children due to this group's care and educational needs (American Psychological Association 2020; Collins et al. 2021;Heggeness et al. 2021). School-age children living in the same home could plausibly have different school modalities. Although we control for child age and number of children in the home, having more than one school-age child likely increased schedule uncertainty and the odds of incompatibilities between COVID concerns and children's school modes. Moreover, it may have led to the need for more care and educational supports (Calarco et al. 2021). Examining the overall care and educational needs in families and the formal and informal supports that they used to buffer stress is an important direction for future research (Yang, Zang, and Calarco 2025).

The measurement of hybrid school modalities is also limited. Although this study addresses hybrid schooling, given the wide variety of hybrid setups and changes in school mode from week to week in some districts, more research is needed to understand the benefits and challenges of varying hybrid school arrangements and shifting school modes for parents and children.

Last, this study is a cross-sectional analysis of the moderating effect of COVID concerns on the association of school mode with stress that cannot make definitive claims about causality among our variables of interest or the long-term consequences of school modality-COVID concern compatibility. Although the interaction estimates that we present are robust to the inclusion of statistical controls in models, results may be a product of unobserved variable bias. Additionally, endogeneity among stress, COVID concern, school mode, and some of our controls (e.g., job flexibility) may bias estimates. To assess causal order among our variables, we ran supplemental analyses limited to panelists who participated in both Wave 1 and 2 of the SPDLC (n = 327). In these models, we regressed parents' stress in fall 2020 on children's school modes in fall 2020 and parents' COVID concerns in spring 2020, with all controls measured in spring 2020, including a measure of parents' stress in spring. The results (see Appendix Table A3 in the online version of the article) were remarkably similar to our main findings presented in Table 2, indicating that net of early pandemic conditions and stress, parents' COVID concerns moderated the association between their children's school modes and parents' stress in fall 2020.

Regarding the long-term consequences of (in) compatibility between school modes and COVID concerns, lower levels of stress among unconcerned parents with children attending school in person may have been short-lived if their family was at substantial risk of contracting COVID and subsequent infections led to dire consequences. A little less than one-half of parents in this study reported concerns about COVID, which appear well founded given that COVID-19 deaths and hospital admissions started to spike in fall 2020 (CDC n.d.-b). Yet COVID risks were unequally distributed in the United States: The socioeconomically disadvantaged were far more likely to contract the disease and die from it than their more advantaged counterparts (Gold et al. 2020). Although research in the first year of the pandemic suggests that concerns over COVID and school reopening policies correlated with exposure risk (Landivar et al. 2022), both were often based on other factors (e.g., political views) because much about the virus was still unknown. It is therefore unclear whether parents' concerns about COVID accurately reflected exposure risk in their children's schools and if this had long-term effects on health and well-being.

Parents' mental health suffered greatly during the COVID-19 pandemic and continues to languish. Although some proffered reopening schools in fall 2020 as a solution to parents' troubles, only a small portion of parents across the United States-those unconcerned about COVID-experienced comparatively lower stress when their children were attending in-person school full-time. By contrast, the plurality of parents who were concerned about COVID exhibited elevated stress levels when their children returned to in-person school compared to continuing with mandated virtual schooling. The push to return to in-person school without adequate parental supports and effort to reduce parental concerns appears to have done more harm to parents' well-being than good. As future pandemics arise, policies to protect the mental health of citizens should focus on understanding and addressing citizens' health concerns, thereby empowering and protecting those most at risk.