The Politics of Re-Opening Schools: Explaining Public Preferences Reopening Schools and Public Compliance with Reopening Orders During the COVID-19 Pandemic

Abstract

Due to the COVID-19 Pandemic, the decision to reopen schools for in-person instruction has become a pressing policy issue. This study examines what overall factors drive public support for schools re-opening in person and whether members of the public are willing to comply with school re-opening decisions based on their own preferences and/or the level of government from which the order comes. Through two rounds of national surveys with an embedded experiment, I find consistent evidence that 1) trust in information from elites - not contact with COVID - best explain preferences for reopening, 2) political ideology and racial and class identification help explain preferences as well, and 3) the President of the United States is best positioned to generate compliance with a school reopening mandate. This study suggests that politics - not public health - drives public support for schools reopening and compliance with government orders to reopen.

What factors determine public preferences for schools reopening in the midst the COVID-19 pandemic? Are people forming preferences based on public health information, social identity, their politics, or a combination of all three? In addition to forming preferences, are people willing to comply with government mandates to reopen schools? To the extent that they are, what are the factors that best determine who is willing to comply versus who is not? These are pressing questions that emerge in through the continued existence of the COVID-19 public health pandemic. Moreover, these are the questions that I will address in this paper.

As the COVID-19 spread rapidly during the early months of 2020, state and local government leaders largely moved at rapid speeds to close school facilities around the United States for the remainder of the 2019-20 academic school year.¹ However, the question of reopening for 2020-21 became far more complicated. While most states did mandate that schools reopen, they varied in the fabric of their guidelines for aspects such as: scheduling, facility requirements, and opt-out plans for individual parents and students (Grossmann et al., 2021). Meanwhile, some states broke from the central guidelines structure and appropriated decision-making control to local districts. As state and districts scrambled, the President and the Department of Education sent out strong messages in favor of schools reopening with limited federal resources to support the process.

The intergovernmental confusion around the school reopening process provides a challenge for public schooling and for democratic accountability (Manna, 2006; Wong, 2008; Cohen & Moffitt, 2010; Henig, 2013). In terms of public schooling, the pandemic raises questions as to who is responsible for COVID-19-related decision-making and implementing special protocol. Meanwhile, the accountability issue raises questions as to whether public concerns are being heard and met. These issues forge together to create questions around the larger issue of public compliance. If the public senses uncertainty from policymakers and/or believe that policymakers don’t have their best interest at heart, there could be incentive to avoid complying with mandates. During a massive public health crisis, non-compliance, when the right policy is in place, is a proverbial highway for further transmission.

This paper examines public preferences towards reopening schools during the COVID-19 pandemic. I uncover the factors that best predict support for reopening schools. I then move to examine the issue of public compliance with orders to reopen schools, and I do so an intergovernmental experimental design. By randomly assigning a hypothetical compliance order to come from the mayor, the governor, or the President of the United States, I assess whether the level of government from which the order comes impacts public compliance. I also perform the same test on a hypothetical stay-at-home order, which provides a comparison to compliance with school reopening mandates. Ultimately, I find evidence that political, social, and economic factors – not contact with the virus itself - influence preferences for reopening schools and compliance with school reopening mandates.

National Crisis and Public Preferences

Public preferences and opinions are a vital part of democracy (Zaller, 1992; Berinsky Adam, 2004). They provide evidence of a public will, which can be a tool for holding representatives accountable. During times of crisis, public opinion becomes particularly important because existential threats like war often require swift action (Berinsky & Druckman, 2007; Berinsky 2009.) In 2020, the U.S., along with countries around the world experienced, experienced a tremendous public health threat, when the spread of the coronavirus (SARS-Cov-2) led to the COVID-19 pandemic. One of the largest policy-related questions that emerged during the early stages of the COVID-19 public health crisis is: in what capacity should schools operate?

Schools demanded significant attention from policymakers because of how central education is to American society. Because of compulsory schooling laws, every child is granted access to public schools, and in most states, laws mandate that parents send children to school until they reach the age of 16 (Katz, 1976). Meanwhile, under the Elementary and Secondary Education Act of 1965 and subsequent reauthorizations, the U.S. Federal Government is for ensuring that Americans have access to schools of a certain quality (Wong, 2008). Because of these state guarantees and federal oversight, governments at the state, local, and federal levels each invest billions of dollars annually in education. As of 2020, U.S. governments spend an estimated $12,612 per student for about 50.7 million children, which amounts to an estimated $639 billion dollars spent per year on public education (U.S. Census 2021). Governments have a vested interest in keep schools open and operating. Governments also have a legal obligation. With the spread of the coronavirus complicating the capacity for schools to operate, the pandemic created a pressing quagmire to which governments needed to respond to swiftly.

Swift policy action under democratic governance, however, should require public support. As such, one of the primary questions I explore in this study is: Did the COVID-19 pandemic lead to clear differences in policy preferences for schools to reopen? The question of public preference, though, opens up the longstanding debates that have driven the study of public opinion. Early research raised doubts as to whether the public can actually form coherent policy preferences and maintain those preferences over time. Instead, individuals form into what Converse, 2006 calls “issue publics” or unofficial groups of people who have a specialized interest or knowledge on specific policy issues. In Converse’s model, people did not form true opinions on policy issues that are low in salience or removed from everyday experience. Subsequent research distinguished between how people form preferences when they have low information versus high information. Zaller (1992) finds evidence that, when citizens are low in policy information, they form policy preferences using “top of the head” judgements. People make these sudden judgements based on cues from information elites. Public opinion research now consistently reiterates the central roles that news sources (Iyengar & Kinder, 1987; Iyengar, 1991; Bartels, 1993; Guardino, 2019) and political parties play (Bartels, 2002; Druckman, 2013) in how members of the general public form policy preferences. The question for this study is: will school reopening policy follow suit?

The Politics of School Reopening During the COVID-19 Pandemic

There is a limited amount of research on the ways in which the COVID-19 pandemic has interacted with the politics of education. However, the research in circulation provides evidence that political factors are operating as strong determinants of both policy responses from governments as well as the formulation of preferences from members of the public. In separate analyses of state reopening plans, Grossmann et al., (2021); Hartney & Finger 2022) find evidence that partisan leanings factored into district policy decision-making. The latter, (Hartney & Finger, 2022), as well as DeAngelis and Makridis. (2020) provide evidence that interest-group organizing lies beneath the partisan politics shaping school reopening. More specifically, they find that the involvement of teacher unions has been shifting policy, particularly in areas run by the Democratic Party. Meanwhile, media coverage has been documenting a similar phenomenon amongst parent groups organizing – many of whom align with ideologically conservative political campaigns.² There is also evidence that partisan control played a significant role in different emergency response policy: stay-at-home orders (Patterson, 2020). Again, while low in volume, there is a consistent trend in that political factors play a heavy role in policy decision-making around school reopening.

We are less sure, however, whether political factors shape the public’s policy preferences for schools to reopen during a public health crisis (or remain closed). In fact, we are largely unsure what factors influence school reopening preferences at all. Most fundamentally, we would expect that factors related to one’s experience with the virus and with schooling during the pandemic would influence policy preferences. But, this may not be the case. Much of the research emerging from political science from the past decade illustrates how policy issues can become politicized (Achen and Bartels 2017; Mason, 2018; Mutz, 2018) and racialized (Tesler & Sears, 2010; Parker & Barreto, 2014; Jardina, 2019) to the point where the primary lens through Americans digest salient policy issues is through partisanship, ideology and/or racial identification. Also, because the COVID-19 Pandemic has disproportionately impacted people of color and low-income households (Chowkwanyun and Reed 2020; Gaynor & Wilson, 2020), we could see differing levels of support for an issue like reopening schools along lines of race and class in particular. So, based on the trends in the literature on school reopening policy as well as the research on policy attitudes, we should expect partisan and ideological leanings to in some way determine policy preferences. Moreover, the latter literatures – combined with real world disparity – also create the room to suspect that social factors, particularly racial identity and economic status, will factor in as well.

It is important to also assess how members of the public respond to the emergency decisions made by public officials. If political factors do play a role in preferences for (or against) school reopening, the source of government orders to take emergency action should matter as well. From a legal standpoint, governors (and state legislatures) have the greatest amount of power to make these types of policy decisions (Briffault, 2005). So, we could expect school reopening orders from governors to generate the most compliance. However, governors have the discretion to reappropriate that power to cities and districts, which might increase the strength of a school reopening order that comes from a school board or a city mayor. Meanwhile, the president, at the federal level, has no legal authority to issue school reopening orders, but they have arguably the largest platform from which to make policy recommendations for states and districts to follow (Canes-Wrone, & Brandice, 2001). Because of the multiple possibilities, I devise an experimental design to determine which level of government actor generates the most compliance and under what conditions. I also assess what additional factors play a role in compliance with school reopening orders, and I examine whether there are consistencies in motivations for school reopening preferences and compliance with school reopening orders.

Research Design

In order to assess questions of preference and compliance, I fielded two iterations of an original survey experiment. I titled the first iteration as “Assessing Opinions of Public Education and Political Participation under COVID-19.” I administer the study through Prolific, which is a market research firm that maintains a network of people who can be recruited for various survey research endeavors. Study one boasts a total of n=1273 responses collected from July 30^th – Aug. 2^nd of 2020. Of the total sample, 83% of respondents are a part of the original nationally representative sample. The remaining 17% are a part of an oversample of low-income people Black and Latinx Americans that I included to better assess potential differences across race and class. The sample in its totality is not nationally representative and the analyses do not include sampling weights. This study is interested in understanding the preferences of vulnerable populations. All respondents completed the survey online in an average of 13 minutes.

The second iteration, Study 2, underwent a similar process. Fielded from Jan. 21^st – Jan. 26^th of 2021, participants agreed to take part in a survey entitled “Reassessing Opinions of Public Education and Political Participation under COVID-19.” Due to university COVID-related financial restrictions, I had to recruit a slightly smaller sample size. Study 2, therefore, features a sample of n = 983. Similar to Study 1, 82% of the respondents for the second iteration form a nationally representative sample. The additional 18% are a part of a low-income Black and Latinx oversample appended to the original 82%. Once again, I administered the survey through Prolific, but in order to ensure that new participants had no prior exposure to the measures and treatments, I screened out participants who took part in Study 1. Surveys were all completed online in an average of 14 minutes.

Measuring Preferences for Reopening Schools

Both iterations of the survey feature a question measuring preferences for school reopening. It is worth nothing that during the time in which Study one was fielded, there was tremendous uncertainty around what school reopening would look like. At that point, the overwhelming majority of schools were closed, and district and state leaders were grappling with the difficult decision of whether to open for in-person instruction or remain physically closed, while operating remotely. There were many factors that created lots of variation in what reopening could look like such as: staggered schedules, non-rotating classrooms, and ventilator mandates. Due to the ambiguity of school reopening policies at this particular time, I employed an open-ended response question to assess preferences for reopening schools. I, then, developed a coding scheme based on what I perceived to be the most linear theme emerging from the responses: reopening with or without safety precautions in place.

The end result is a categorical variable that ranges from 0–1. There are a total of five categories: respondents in support of reopening (coded as 0), leaning towards a preference that schools reopen (coded as 0.25), expressing indecision (coded as 0.5), leaning towards opposition (coded as 0.75), and respondents in firm opposition (coded as 1). The nuance of the responses dictated the need for the three categories in-between firm support and opposition. Within the qualitative data, I find a significant number of conditional responses. For example, a White male higher-income respondent stated:

“I think schools should reopen only if they can do it safely, while minimizing the risk of transmission to teachers and students.”

Similarly, a middle-aged Latinx mother who opted to take the survey in Spanish responded with:

“Dependiendo de la gravedad de la situación del momento [Depending on the severity of the situation at the time].

Meanwhile, other study participants offered straight-forward answers about the complexity of the issue. “I am grateful that I don’t have to make the decision to re-open the schools; it is a tough call,” wrote an older White male Republican from the sample. A young Black male revealed that, “It’s stressful and mentally taxing. [It] depresses me.” The categorizing that I employ in this design helps do capture the nuance in the participants’ responses.

For Study 2, I use a more standard measure. By January 2021, the school reopening debate had reached a higher level of salience. Media outlets were widely discussing the implications of schools opening for the Spring 2021 semester and returning to in-person instructions with (and even without) safety precautions. Because of the high salience, I measure school reopening preferences in 2021 through a close-ended response question. Survey respondents were shown the following language:

“A number of schools around the U.S. are attempting to reopen for in-person instruction. Do you agree or disagree with schools reopening for in-person instruction?”

Respondents are presented with a set of response options on a likert scale from strongly agree to strongly disagree.

Compliance Experiment Design

In addition to base preferences for reopening schools, I am particularly interested in the effect of orders from government leaders to reopen schools. The issue of reopening schools, though, faces the intergovernmental dilemma. State governors and legislatures have most legal authority to determine the operation status of schools, but state leaders also have power to abdicate decisions to the local level. Meanwhile, the President of the United States, Secretary of Education, and members of Congress all have, arguably, a larger platform for disseminating political and policy-related messages. So, when it comes to government orders to reopen schools, which level of government actor generates the most compliance: local, state, or federal?

I address this question by embedding an experiment into both Study one and Study 2. Through the experiment, I take the idea of government officials offering an order for schools to reopen, and I manipulate the level of government from which that order comes. I then measure respondents’ self-reported willingness comply with the order. More specifically, during the survey, I randomly assign a hypothetical school reopening mandate statement prior to assessing their willingness to comply with that hypothetical mandate. Table 1 displays the actual language from the treatments. The measure of compliance is structured on a Likert scale ranging from strongly agree to strongly disagree. Both studies use the exact same language.

Table 1.

Compliance Experiment(s) Language.

Experiment Elements	Language
School reopening mandate
Background	There is uncertainty around whether public schools should re-open at all before the majority of Americans receive the COVID-19 vaccine. However, if the number of COVID-19 cases does not increase significantly over the next few weeks, schools may be ordered to re-open. Indicate whether you agree or disagree with the next statement
Mayor treatment	If the mayor of my city eventually orders that schools open, I will comply with that order
Governor treatment	If the governor of my state eventually orders that schools open, I will comply with that order
President treatment	If president [trump/Biden] eventually orders that schools open, I will comply with that order
Stay-at-home order
Background	This is uncertainty around whether public spaces and businesses should remain open. If the number coronavirus cases significantly increase, stay-home orders could be re-issued. Indicate whether you agree or disagree with the statement in the next question
Mayor treatment	If the mayor of your city eventually orders everyone to stay-at-home, I will be willing to comply with that order immediately. No questions asked
Governor treatment	If the governor of my state eventually orders everyone to stay-at-home, I will be willing to comply with that order immediately. No questions asked
President treatment	If the president of the United States eventually orders everyone to stay-at-home, I will be willing to comply with that order immediately. No questions asked

In addition to the school reopening mandate compliance experiment, I include a test of compliance with emergency stay-at-home orders. While the primary focus of this study is the former, I employ an assessment of stay-at-home orders to provide a non-education-related comparison to compliance with school reopening mandates. I, therefore, randomly assign respondents to receive a hypothetical stay-at-home order from either the mayor, governor, or president. The specific language for the stay-at-home order experiment is also displayed on Table 1, and I use the exact same response options as for the school reopening mandate compliance experiment.

Explaining Opposition to Reopening Schools

Before venturing into the effects of the compliance messages, it is important understand the raw preferences for reopening schools. From there, we can begin to explore the factors that are most closely associated with having a particular preference. In terms of raw preference, though, it is clear that, as expected, assessing preferences at different points during the pandemic yielded a different distribution of responses. There were much stronger preferences for schools to remain closed amongst the participants of Study 1 (Fall 2020) compared with participants of Study 2 (Spring 2021). As Figure 1 demonstrates, almost two-thirds of the sample (66%) firmly opposed reopening schools, but for Study two only 47% of respondents expressed opposition. In juxtaposition, support for reopening schools increases during this period; moving from 23% support up to 41%.

Figure 1.

Preferences for Reopening Schools at the Beginning of Fall 2020 versus Spring 2021.

Note. Line plots show preferences rates for reopening schools for in-person instruction by policy position. All estimates are averages.

The question, though, is what factors seem to be driving these preferences? Moreover, are the factors consistent over time, or are they different during different phases of the pandemic? It makes sense that overall preferences would change as new treatments emerge and hospitalization and death rates decline, but that may not necessarily change the motivation for the preferences. This could be particularly true if Americans are responding to changes in public health information or responding based on their on level of contact with the actual virus itself. It could also be true, if Americans are forming preferences based on messaging from political elites. In other words, it its important to explore whether the change in the distribution of preferences is a function of changes in the factors motivating the preferences or changes in the conditions of the factors that motivate the preferences. More simply: are people responding to different factors over time or the same factors that just so happen to be experiencing different conditions over time?

The answer appears to be a combination of both. Table 2 provides OLS regression estimates for the same model attempting to explain preferences for school reopening in Study August 2020 compared with January 2021.³ For August 2020 (see Table 2; August 2020 column), contact with the actual virus (or even knowing someone who has) was statistically unrelated to school reopening preferences. Instead, trust in COVID-related information from two sources – the CDC or the President Trump - emerge as strong correlates.

Table 2.

Modeling Preferences for Reopening Schools.

	Support for Reopening Schools
	Aug 2020	Jan 2021
Covid contact scale	−0.001	0.006
	0.058	0.039
Impacted by learning loss	0.091*	0.063
	0.038	0.035
Trust covid info (news)	−0.081	0.311**
	0.044	0.037
Trust covid info (CDC)	−0.118***	−0.019
	0.033	0.027
Trust covid info (president)	0.300***	0.111**
	0.038	0.036
Trust covid info (governor)	−0.007	0.001
	0.034	0.063
Trust covid info (mayor)	−0.079	0.086
	0.042	0.063
District satisfaction	−0.034	0.049
	0.042	0.040
Parent (school-aged child)	−0.002	0.034
	0.022	0.020
Black	−0.076**	−0.067***
	0.029	0.024
Latinx	−0.094**	−0.070*
	0.032	0.029
Asian	−0.119**	−0.128***
	0.043	0.037
Other	−0.026	−0.092
	0.071	0.079
Female	−0.012	−0.038*
	0.020	0.017
Democrat	−0.058	−0.010
	0.031	0.028
Independent	−0.031	0.042
	0.029	0.026
Ideology (1 = conservative)	0.201***	0.190***
	0.043	0.039
High income	0.064*	0.066**
	0.029	0.026
Middle income	0.038	0.036
	0.023	0.021
N	1273	983

***p < 0.001; **p < 0.01; *p < 0.05.

Note. Estimates are derived from ordinary least squares (OLS) multivariate regression models.

Respondents who expressed high levels of trust in COVID information from the CDC were 12 percentage points less likely to support school reopening, while individuals trusting information from President Trump were 30 percentage points more likely to support reopening schools for Fall 2020. Both estimates are statistically significant at the 99% confidence-level. CDC and President Trump were largely at odds during this period, and their division mapped onto school reopening preferences. In early July of 2020, President Trump tweeted, “I disagree with the @CDCgov on their very tough and expensive guidelines for opening schools.”⁴ While he went on to acknowledge that, “[the CDC] wants them open,” He asserted that, “they are asking schools to do very impractical things. I will be meeting with them!!!” The political lines were clear.

If August 2020 is a show of the division between the CDC and President Trump, January 2021 shows an immediate congruence between newly inaugurated President Biden and news media coverage of COVID. Biden makes the push to reopen schools for in-person instruction one of the primary goals of his first 100 days in office. The difference, though, is that Biden consistently emphasizes the need for health and safety restrictions as a central part of his plan. The American Rescue Plan Act that he pushes for and signs into law allocated millions of dollars to schools for ventilators and other health and safety equipment. Meanwhile, reports from mainstream news outlets circulated more updated information on the weaker impact of COVID-19 on children and the negative impact of COVID-19 on the American economy. Biden’s plan was reflective of a different information environment.

The data reflect this. Biden took a more aggressive stance on reopening than the CDC.⁵ Suddenly, trust in information from the CDC becomes an unrelated predictor of support for reopening schools. Meanwhile, trust in information from the news becomes be strongest predictor. Study two respondents were 31% points more likely to support school reopening, if they trusted the information from the news. Meanwhile, trust in information from the president remains leading factor in support for reopening schools as well; those trusting newly inaugurated President Biden were 11% points more likely to support school reopening.

It is important to note that the strength and direction of the correlations, though, stand apart from the pervasiveness of the factors themselves. For instance, trust in information from the news media becomes positively correlated with school reopening preferences (and the most robust factor predicting preferences for reopening) at the beginning of 2021. However, as Figure 2 shows, overall trust in the news coverage is extremely low by 2021 (20%); it declines by 35% points from August 2020 to January 2021. Meanwhile, trust in the CDC also dips over time, but only by 14% points. Moreover, in 2020 – when CDC trust was a strong predictor of opposition to school reopening – an overwhelming 81% of respondents expressed trust in their information.

Figure 2.

Trust in COVID-19 Information By Source (President, News, or CDC). Note. Line plots show trust in COVID-19 information by different sources. All estimates are averages.

So, as trust declines for both the CDC and the news media, movement in statistical significance from the former to the latter suggests that Americans changed their rational for supporting reopening. It is a transition from rejecting the information from the CDC to embracing the optimism of a return to normal life that people seem to have pieced together from media coverage of the pandemic. It is worth noting, though, that political conservatives expressed trust in news coverage of COVID-19 at a much higher rate (50%) than liberals (7%). Thus, it seems most likely that the media coverage that favors reopening was likely predominantly happening on conservative-leaning outlets. As such, trust in news as a motivating factor for supporting the reopening of schools appears to largely be a phenomenon amongst political conservatives.

A unique phenomenon occurs with trust in information from the president, however. While trust in information from the news media and CDC decline over time, trust in COVID information from the president increases by 11% points between August and January. The glaring difference, here, is that this increase happens as we see an administrative change, with the White House transitioning from Donald Trump to Joseph Biden. So, the increase could be either a “Biden bump” or the removal of a “Trump penalty.” Either way, not only do we see an increase in trust, but despite the administrative decision, the relationship direction remains the same (positively correlated with preferences for reopening schools). Moreover, trust in COVID information from the president is the only trust-related factor that reaches statistical significance in regression models for both iterations of the study. Trust in the president’s COVID messaging, while in low absolute terms, was a consistent and growing factor.

While the role of trust in information from the president was consistent across the two studies, so too were the social and political backgrounds of the people who supported school reopening. For instance, political conservatives and individuals of high-income households are statistically more likely to support school reopening. Meanwhile, people of color – Black Americans, Latinx Americans, and Asian Americans – and respondents identifying as women are all statistically less likely to oppose school reopening. Moreover, the magnitude of each statistical relationship remains relatively the same across the two waves of data collection. This consistency along the dimensions of political ideology, socioeconomic status, and racial and gender identification underscores an important notion. While we saw an overall decline in opposition to school reopening, people largely continued to hold strong preferences for schools to reopen (or remain closed) based on how they processed information through their social, economic, or political lenses. The latter dimension seems to be what changed most, with trust in Biden and Trump performing the same way, and this likely explains the bulk of the overall change support for school reopening over time.

Who Will Comply with School Reopening Mandates?

As correlational analyses reveal, cues from elites – particularly the President of the United States, CDC experts, and the news media – factor heavily into preferences for schools to reopen for in-person instruction. The response, or lack thereof, to political elites is particularly important, though, as they have the most say over how schools reopen and under what conditions. In fact, it is important to reemphasize that, for the vast majority of situations, the state governor (and the legislatures) carry the most legal authority to execute an order to reopen or close schools. Governors, however, also have the ability to leave decisions to be made at the discretion of the mayors, city councils, and school boards. The president, meanwhile, has virtually no legal authority on the matter, but they have the largest platform through which to have their recommendations heard by the public. Recall that, in addition to the president’s mega platform, trust in information from the president was the only consistently important factor driving preferences for reopening schools. Conversely, trust in information from governors and mayors, showed no statistical relationship, which indicates our ability to reject the null hypothesis.

This brings us to the question of compliance. Not factoring state and local officials into how one forms preferences for reopening is one phenomenon, but are people responsive to their orders? More specifically, as leaders at all three levels of government (local, state, federal) have to revisit the possibility of mandates for schools to reopen: Are Americans responsive to such mandates? If so, to whom are they most responsive? Which level of government elite generates the highest level of compliance? To address these questions, I turn to the compliance experiment embedded within the survey. Again, I present a hypothetical order for schools to reopen (in August 2020) and again (in January 2021), and I randomize whether that hypothetical order comes from the mayor of the city, their state governor, or the President of the United States.

The evidence from the experiments suggest that the president’s messaging matters tremendously. Respondents in the first round of the study from Fall 2020 were the least likely to comply with school reopening, when the order was attributed to President Trump (See Figure 3). While the governor received a 55% compliance rate and mayors a 48% compliance rate, only 37% of respondents would comply with the hypothetical order from the President of the United States. For the second round of the study, the results flip. Respondents who are assigned the compliance order from President Biden are the most likely (65% rate) to comply with a hypothetical order for schools to reopen. Meanwhile, the compliance rate for governors remained the same (55%) and the compliance rate for mayors eked up to 56%. How survey participants responded to the mandate from the president, though – whether by acceptance or rejection – seems to be what dictated their willingness to comply.

Figure 3.

Willingness to comply with school reopening order over time by level of political.

Note. Line plots show the school reopening compliance rates by treatment group. All estimates are averages.

Do the difference-in-difference estimates hold when considering other factors? Table 2 displays regression results for models estimating compliance with school reopening orders as a function of the mandate treatments plus the factors that I used to model the preferences for reopening schools. Consistent with the difference-in-difference estimates, the mandate from the president is a strong predictor in the regression models as well. Individuals who received the president mandate treatment were statistically less likely to comply with a school reopening order in 2020. Meanwhile, those who received the treatment were more likely to comply in 2021. Both estimates are statistically significant at least the 95% confidence level.

So, is compliance purely about reactions to presidential figures? The additional estimates that surface from the regression models suggest that the story is more complicated than that. There are some factors that are robust predictors in one wave of the survey only. For instance, respondents during August 2020 were less likely to comply with any government mandate if they held high trust in COVID information from the CDC. Conversely, respondents who trusted information from President Trump were much more likely to comply with a mandate. Neither factor reaches statistical significance in January 2021. In short, the same transition that we see for preferences for reopening schools, this evolution from the draconian battle between President Trump and the CDC to trust (or distrust) in the news media guiding preferences, surfaces within intentions to comply with a reopening mandate.

Social, economic and political factors, once again, play a role. The importance of those factors, however, is not consistent across both waves. In 2020, Democrats and Independents were statistically less likely to comply with the reopening mandate. Meanwhile, in 2021 respondents identifying as either Black or female became less likely to comply. Instead, it was the individuals of high-income households who were more willing to comply in 2021.

There also three factors that consistently serve as robust predictors across the two waves: learning loss, school district satisfaction, and political ideology. Survey participants in 2020, who expressed the greatest fear of learning loss for students of families like theirs, were about 15% points more likely to comply than those who expressed the least amount of concern. That difference dips to about 10% points in 2021. Respondents who were satisfied with their school district overall were about 9% points more likely to comply with a reopening mandate in 2020 and about 15% points more likely to do so in 2021. Meanwhile, extreme political conservatives were about 8% points more likely to comply in 2020 and about 10% points more likely to comply in 2021. This suggests that the persisting divide in Americans’ willingness to comply with school reopening orders, to some degree, comes down to their spatial position on the ideological spectrum and whether they live in a school district that they consider to be of high quality. However, it is also a function of how much they anticipate kids falling behind academically.

What is perhaps most compelling is that these more school and education-related factors are largely silent within the compliance results (and the school reopening results as well). For instance, while learning loss concern factored into compliance, it was only a statistically significant predictor of preferences for reopening schools in 2020 and only at the 90% confidence level (See Table 2). Similarly, parents of school-age children were statistically no more or less likely to support, and parental status only factors into compliance in 2020 (an estimate that is, again, only statistically significant at the 90% confidence level). Moreover, actual contact with the virus largely has no bearing on preferences for schools to reopen nor people’s willingness to comply with a reopening order. In other words, compliance with school reopening orders (and preferences for schools to reopen) seem to be only marginally driven by the lived experiences of children and families Table 3.

Table 3.

Modeling Compliance with School Reopen Orders.

	Compliance with School Reopen Order
	Aug 2020	Jan 2021
President treatment	−0.128***	0.066**
	0.020	0.020
Governor treatment	0.017	−0.019
	0.019	0.020
Covid contact scale	0.018	0.046
	0.047	0.037
Learning loss concern	0.153***	0.096**
	0.030	0.034
Trust covid info – News media	−0.002	0.087**
	0.035	0.035
Trust covid info – CDC	−0.076**	0.042
	0.027	0.026
Trust covid info – President	0.343***	0.065
	0.030	0.035
Trust covid info – _governor	0.048	−0.025
	0.027	0.060
Trust covid info – Mayor	−0.012	0.117
	0.034	0.060
District satisfaction	0.087**	0.148***
	0.034	0.038
Parent (school-aged child)	0.037*	0.002
	0.018	0.020
Black	−0.038	−0.048*
	0.023	0.023
Latinx	−0.037	−0.046
	0.026	0.027
Asian	−0.048	−0.006
	0.035	0.035
Other	−0.086	−0.066
	0.057	0.075
Female	−0.011	−0.044**
	0.016	0.017
Democrat	−0.079**	−0.028
	0.025	0.027
Independent	−0.038	−0.027
	0.023	0.025
Ideology	0.081*	0.097*
	0.034	0.038
High-income	0.027	0.067**
	0.023	0.025
Mid-income	0.008	0.036
	0.018	0.020
N	1273	983

***p < 0.001; **p < 0.01; *p < 0.05.

Note. Estimates are derived from ordinary least squares (OLS) multivariate regression models.

Is Compliance With School Reopening Mandates Unique?

The evidence on compliance with school reopening mandates suggests that public preferences seem to be largely driven political factors. However, it is important to know the extent to which the issue of school reopening is unique to other emergency response policies. In order to provide a comparison, I incorporated an additional survey experiment that focuses on the issue of stay-at-home orders. Just as with the school reopening mandate, I assign the mandate for the stay-home order to come from the mayor, the governor, or the president.

There are two main distinctions that emerge between the stay-home order mandate preference assessment and the school reopening mandate. The major difference is that compliance with stay-at-home orders is significantly high across the board. Figure 4 shows the exact estimates. Unlike school reopening mandate compliance, over 80% of respondents were willing to comply with stay-at-home orders regardless of which leader the order came from. Moreover, there is very little change in compliance over time. The biggest change is an extremely small 3% point drop in compliance with an order from mayor between 2020 and 2021. Regression results (See Appendix) Table A1 confirm that the differences in whom the mandate comes from does not reach statistical significance in either survey effort. Americans are much more resistant to school reopening mandates than stay-at-home orders.

Figure 4.

Willingness to comply with stay-at-home orders over time by level of political.

Note. Line plots show the stay-at-home order compliance rates by treatment group. All estimates are averages.

Table A1.

Modeling Compliance with Stay-at-home Orders for Fall 2020 versus Spring 2021.

	Fall 2020	Spring 2021
President treatment	−0.026	−0.022
	0.018	0.016
Governor treatment	0.014	−0.007
	0.018	0.016
COVID contact scale	0.034	0.030
	0.033	0.039
Learning loss concern	−0.005	−0.006
	0.030	0.025
Trust COVID info – News media	−0.395***	0.138***
	0.031	0.029
Trust COVID info - CDC	0.153***	0.152***
	0.023	0.022
Trust COVID info - president	−0.011	−0.202***
	0.031	0.025
Trust COVID - governor	0.136*	0.069**
	0.053	0.022
Trust COVID info - mayor	−0.063	0.074**
	0.053	0.028
District satisfaction	0.050	0.015
	0.034	0.028
Parent (school-aged child)	0.001	0.020
	0.017	0.015
Black	0.036	−0.001
	0.020	0.019
Latinx	0.045	0.037
	0.024	0.021
Asian	0.060	0.004
	0.031	0.029
Other	−0.004	0.028
	0.067	0.048
Female	0.049***	0.019
	0.015	0.013
Democrat	0.031	0.043*
	0.024	0.021
Independent	−0.042	0.000
	0.022	0.019
Ideology	−0.141***	−0.042
	0.033	0.028
High-income	−0.036	−0.022
	0.022	0.019
Mid-income	−0.018	−0.033*
	0.017	0.015

***p < 0.001; **p < 0.01; *p < 0.05.

Note. Estimates are derived from ordinary least squares (OLS) multivariate regression models.

The difference in issue compliance is likely driven by the differences in proximity to an enhanced health environment. Stay-at-home orders are a tool to prevent virus spread, while school reopening mandates make the public vulnerable. Compliance is much higher for the order that enhances safety than for the issue that injects more harm. But, if this were the sole explanation, Americans would comply with school reopening mandates at the inverted rate at which they comply with school reopening orders. For example, if 86% of respondents are willing to comply with stay-at-home orders (purely because of its emphasis on safety), we should expect for 14% of respondents to comply with school reopening mandates (given its vulnerability to harm). Instead, we see school reopening compliance mandate compliance rates north of 50%. Compliance is about more than the public health implications.

I argue that the differences between compliance with stay-at-home orders and school reopening mandates also show us that the later is uniquely political. While not a total inverse in compliance rates, there were still major differences in compliance with school reopening mandates compared to compliance with stay-at-home orders. The difference is particularly pronounced when assessing the reactions to orders/mandates from the president. Consider the way in which President Trump politicized reopening schools,⁶ while taking a softer stance on stay-at-home orders.⁷ To the former, the president pushed states and districts to take action, while on the latter the White House quietly issued guidelines and openly left the decisions to state leaders. Without the national media spotlight (and hard public stance), public compliance reaches high levels and maintains them.

Discussion

The COVID-19 pandemic has raised important questions for the politics of education. The ubiquity of the pandemic, in particular, thrusted school reopening onto the forefront of major policy debates. As the politics of these decisions grow more contentious on the surface, it is imperative to better understand what factors seem to drive those preferences. Through analysis of two waves of survey data collected at the beginning of two academic semesters, my results reveal a somewhat complex story. Opposition to school reopening for in-person instruction appears to decline, as the U.S. makes progress on combatting the virus. As a vaccine dissemination plan develops and both hospitalization and death rates decline, support for reopening schools seems to increase on average.

The persisting public divide on school reopening, though, is largely unrelated to factors connected to the lived experience of school during the pandemic. Having contracted the virus or knowing someone who has did not make respondents any more or less likely to support school reopening. Neither being a parent of a school-aged child or expressing concern that kids of families like theirs will be a victim of learning loss seemed to determine preferences either. One’s actual stake in the outcome seems not to matter much at all.

Instead, politics reigns supreme. People supportive of school reopening seem to be largely driven by political conservatism. Beyond political ideology, the political lens through which Americans digest COVID-related information plays an important role. Those who distrusted the CDC (in 2020) were more likely to support reopening schools, and so too were individuals who trusted information from the news media. However, trust in the CDC and the news media, respectively, differ tremendously across the ideological divide, with liberals more trusting of the CDC and conservatives more trusting of news.⁸

Trust in the president provides the only resemblance of continuity. Those who trusted President Trump were likely to favor reopening in 2020, while those trusting President Biden were more likely to support reopening in 2021. There is a critical difference in entrusting these leaders. President Trump readily trafficked in misinformation about the COVID-19 virus and minimized the severity of the pandemic. President Biden relied on the expertise of the public health experts to guide his stances and recommendations. Trust in these two leaders, while leading to the same statistical results, are not the same in practice.

The results of the compliance experiment illustrate the divergence in how the public responded to the different presidents. When the hypothetical mandate to reopen schools was attributed to President Trump, barely more than a third of respondents expressed a willingness to comply with the order. The results were the exact opposite for President Biden; all but about a third of respondents were willing to comply with the order, when linked to him. Ironically, the president lacks the power to make such a mandate, but the results suggest that messaging from the president – when public trust in place – can play a critical role in getting the public to take emergency action. Moreover, the fact the federal executive treatment was so effective for school reopening mandates, while indeterminate of compliance with stay-at-home orders, suggests that school reopening, as a salient education policy, was uniquely politicized during the pandemic.

The latter point underscores the upshot of the results of this study. School reopening preferences are extremely political. Naturally, compliance with the idea of a school reopening mandate is extremely political as well. Americans seem to literally follow the [political] leaders, whom they trust. An unfortunate consequence is that ideological disagreement clouds the need to be responsive to the needs of our most vulnerable kids. People of color and individuals of lower-income households were consistent less likely to support reopening schools for in-person instruction. We also know that these are the communities that have been disproportionately impacted by the virus. When the needs of low-income kids, especially low-income kids of color, does not enter the purview of the political debate, the government response is likely to omit their concerns.

Author Biography

Jonathan E Collins is an assistant professor of education, political science, and International and Public Affairs.

ORCID iD

Jonathan E Collins https://orcid.org/0000-0002-7296-1340