Content uploaded by Daniel A. Collier
Author content
All content in this area was uploaded by Daniel A. Collier on Mar 30, 2021
Content may be subject to copyright.
C2i Working Paper Series | No. 210101 | February 2021
We Want You Back: Uncovering the Influences on
In-Person Instructional Operations in Fall 2020
Daniel A. Collier 1,2, Dan Fitzpatrick 3, Madison Dell 4, Sam Snideman 5,
Christopher R. Marsicano 6, and Robert Kelchen 7
Abstract: Institutional responses to COVID-19 are a topic of much concern.
Emergent research has suggested that politics and polarization was more strongly
linked, than was COVID-19, to institutions engaging in-person instruction for Fall
2020. This study used Structural Equation Modeling to test this trend. Based upon
political polarization and dependency, we used data from the College Crisis
Initiative (C2i), to test how state and county sociopolitical features, state and
county COVID-19 rates, and state revenue losses influenced in-person instruction
by September 9th, 2020. The accepted overall model, developed using the full
sample, suggested that County Sociopolitical Features (r=.13) were the stronger
influence on the decision, followed by Pandemic Severity (r=-.10) and State
Sociopolitical Features (r=.09). In recognizing that institutional subsectors may
be uniquely sensitive to these factors we tested our models using the following
subgroups: 4-year public, 4-year private, and 2-year public institutions. State
Sociopolitical Features (r=.17) were the only significant influence on 4-year public
institutions. Whereas, 4-year private and 2-year public institution decisions were
influenced by both State- and County-Sociopolitical Features – these features
were respectively 2x and 3x stronger than were state features. Finally, Pandemic
Severity (r=-.09) only influenced 4-year private institutional decisions to engage
in-person instruction but to a weaker degree than both levels of sociopolitical
features. Overall, models suggest that COVID-19 was not a consistently strong
factor for institutions when deciding in-person instruction and that sociopolitical
features were more influential, including for 4-year private institutions – which
illustrates a propensity towards remaining in favor with sociopolitical “in-groups.”
Keywords: COVID-19; Politics; Dependency; Institutional Decision-Making
JEL Codes: I19, I23
1. W.E. Upjohn Institute for Employment Research
2. University of North Georgia
3. University of Michigan
4. Stanford University
5. Ball State University
6. Davidson College
7. Seton Hall University
C2i Working Papers have
not been peer-reviewed.
Do not cite or quote
without author permission.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
C2i Working Paper Series | No. 202101 | Collier et al. | February 2021
i
[ Page left intentionally blank.]
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
C2i Working Paper Series | No. 202101 | Collier et al. | February 2021
ii
Working Paper Disclaimer and Correspondence Information:
C2i Working Papers have not been peer-reviewed and are available for comment
and discussion only. Do not cite or quote without author permission. Written
correspondence for the purpose of comment, discussion, questions, or permission
to cite or quote this working paper should be addressed to:
Daniel A. Collier, Ph.D.
3636 Lark Drive
Kalamazoo, MI 49008
collier@upjohn.org
Additional Author Information:
Daniel A. Collier, Ph.D.
Kalamazoo, MI, USA
@Dcollier74
Research Associate
W.E. Upjohn Institute for
Employment Research
Part-Time Faculty
University of North Georgia
Dan Fitzpatrick, Ph.D.
Ann Arbor, MI, USA
@FitzEdPolicy
Research and Assessment Specialist
University of Michigan
Madison Dell, M.A.
Palo Alto, CA, USA
@Madisonmdell
Ph.D. Student in
Economics of Education
College of Education,
Stanford University
Sam Snideman, Ed.D.
Muncie, IN, USA
@sam_snideman
Director of Government Relations
Ball State University
Christopher R. Marsicano, Ph.D.
Davidson, NC and Nashville, TN, USA
@ChrisMarsicano
Assistant Professor of the Practice
of Higher Education
Davidson College
Founding Director
The College Crisis Initiative (C2i) at
Davidson College
Robert Kelchen, Ph.D.
South Orange, NJ, USA
@rkelchen
Associate Professor of
Higher Education
Seton Hall University
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
C2i Working Paper Series | No. 202101 | Collier et al. | February 2021
iii
[ Page left intentionally blank.]
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 1
Abstract
Institutional responses to COVID-19 are a topic of much concern. Emergent research has
suggested that politics and polarization was more strongly linked, than was COVID-19, to
institutions engaging in-person instruction for Fall 2020. This study used Structural Equation
Modeling to test this trend. Based upon political polarization and dependency, we used data from
the College Crisis Initiative (C2i), to test how state and county sociopolitical features, state and
county COVID-19 rates, and state revenue losses influenced in-person instruction by September
9th, 2020. The accepted overall model, developed using the full sample, suggested that County
Sociopolitical Features (r=.13) were the stronger influence on the decision, followed by
Pandemic Severity (r=-.10) and State Sociopolitical Features (r=.09). In recognizing that
institutional subsectors may be uniquely sensitive to these factors we tested our models using the
following subgroups: 4-year public, 4-year private, and 2-year public institutions. State
Sociopolitical Features (r=.17) were the only significant influence on 4-year public institutions.
Whereas, 4-year private and 2-year public institution decisions were influenced by both State-
and County-Sociopolitical Features – these features were respectively 2x and 3x stronger than
were state features. Finally, Pandemic Severity (r=-.09) only influenced 4-year private
institutional decisions to engage in-person instruction but to a weaker degree than both levels of
sociopolitical features. Overall, models suggests that COVID-19 was not a consistently strong
factor for institutions when deciding in-person instruction and that sociopolitical features were
more influential, including for 4-year private institutions – which illustrates a propensity towards
remaining in favor with sociopolitical “in-groups.”
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 2
We Want You Back: Uncovering the Influences on
In-Person Instructional Operations in Fall 2020
The COVID-19 pandemic has put American higher education institutions in extremely
challenging positions. As COVID-19 became a concern in Spring 2020, many institutions
quickly pivoted from in-person to online instruction – a justified adaptation to a generally
unknown threat. As summer progressed, institutional decision-makers were balancing the
severity of the pandemic, financial constraints, and sociopolitical pressures when determining
whether Fall 2020 instruction should primarily be in-person, or not. By October, 27% of
institutions had engaged in-person instruction (The Chronicle of Higher Education, 2020).
Emergent research has suggested that political power structures and budget concerns were more
strongly linked to this decision, than was the severity of the pandemic (Collier et al., 2020b;
Felson & Adamczyk, 2021) – a response not unique to higher education when reacting to
COVID-19 (Corder, et al., 2020; Holman et al., 2020).
Unfortunately, this decision came with grim consequences; the University of North
Carolina at Chapel Hill became the public face of poor planning, ceasing in-person instruction
after ten days but producing multiple clusters of outbreaks and the infection of over 1,200
students and employees by September 2020 (Boyd, 2020). By early October, few institutions
made mode of instruction changes, unlike Chapel Hill. Those institutions listed as in-person in
October were tied to 136,838 cases at those institutions by December (The New York Times,
2020b). With imperfect testing programs, that number is likely an undercount. Traditional-aged
students are unlikely to die from COVID-19, which was a selling point for in-person instruction
by some administrators (Lederman, 2020). However, the direct harm this decision caused is
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 3
becoming clearer as evidence accumulates regarding effects on students, staff, and communities.
Even among students with mild initial cases, long-term health effects (e.g. diminished lung
capacity) are emerging (Singh, 2021). Institutions with in-person operation in total have resulted
in at-least 90 staff and faculty deaths from Fall to Winter 2020 (The New York Times, 2020b).
In aggregate, during the Fall semester the choice to engage in-person instruction has resulted in a
daily increase of over 6,500 COVID-19 cases (Andersen et al., 2021). This spread from
campuses to their co-located communities placed vulnerable populations (e.g. those in nursing
homes) at risk; genetic sequencing confirms that campus infections preceded off-campus
fatalities (Richmond et al., 2020).
Given that public trust in higher education has already experienced erosion (Doherty et
al., 2017), if research continues to confirm that higher education played a role in the spread of
COVID-19, recovering public trust may be at-best difficult. In light of both the health-related
ramifications of in-person instruction and the descriptive findings indicating that colleges did not
consider pandemic severity in selecting mode of operation; theoretically-based research using
stronger methods is critical. Before reaching conclusions that may shape public trust, it is also
crucial to understand any differences in how post-secondary sectors reached their decisions.
Leaning on the theoretical underpinning of political polarization (Iyengar et al., 2012; 2019) and
resource dependency (Fowles, 2014), this paper tests prior descriptive findings (Collier et al.,
2020b) using structural equation modeling (SEM) to examine the influences of state and county
sociopolitical features, state revenue declines, and state and county rates of COVID on the
outcome of higher education institutions deciding to reopen primarily in-person for the Fall 2020
semester.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 4
1. How did COVID-19 cases, state and county sociopolitical features, and institutional
characteristics influence the decision to be in-person for Fall 2020?
2. Given different dependencies and political motivations, we further asked how did the
same factors differentially influence decisions at 4-year public, 4-year private, and 2-year
public institutions?
Review of the Literature
Emergent Research on COVID-19 and Campus Re-Opening, Impacts, and Mitigation
The limited research on the impact of COVID-19 on college campus reopening
operations broadly fits into three categories: institutional decision-making processes, health, and
mitigation for education. The first category examined factors including isomorphism, politics,
and revenue. For example, Marsicano et al. (2020) posited that institutions exhibited isomorphic
behavior in shifting to online instruction in Spring 2020. For the Fall 2020 semester, Collier et al.
(2020b) found that a Republican governor was associated with a lower chance of choosing an
online-only mode of instruction, and a Republican legislature was associated with a greater
chance of an institution having in-person instruction – and that political leadership was linked to
the decisions of 4-year and 2-year public institutions, as well as 4-year private institutions. A
trend supported by Felson and Adamczyck (2021). Castiello and Whatley (2021) found a
revenue motive for COVID-19 decision-making: the number of international students enrolled at
a given campus positively predicts that institution’s decision to teach in-person for Fall 2020.
The second category focused on the public health components of campus reopening
plans. Several studies have shown that student mobility (on- and off-campus) increases the
number of COVID-19 cases on campus and in the surrounding community. For example, when
students returned to campus for a hybrid or in-person semester, they led to an aggregate
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 5
nationwide increase of 6,500 cases per day in the early days of the semester (Andersen et al.,
2021). Furthermore, Spring Break travel in Spring 2020 spread the disease on college campuses
and beyond (Mangram and Niekamp, 2020). Through genomic sequencing, Richmond et al.
(2020) determined that students spread the disease to the residents of campus towns, resulting in
increased morbidity and deaths for vulnerable populations.
The third arm of COVID-19 in higher education literature focuses on mitigation
strategies to support or allow on-campus learning environments. Paltiel et al. (2020) estimated
that repeated testing of students could reduce COVID-19 cases to manageable levels. Some
institutions – such as Duke University - have engaged in consistent testing, managing tens of
thousands of tests during the Fall semester, which has helped these institutions control spread
(Denny et al., 2020). However, robust testing was not the norm on college campuses in Fall 2020
(Marsicano et al., 2021). For every well-resourced institution like the University of Illinois at
Urbana/Champaign or Duke University – who have averaged thousands of tests per day - are
several less-resourced institutions who tested far fewer students.
Political Polarization and Partisanship in the U.S.
Social Identity and COVID. Prior to and while the COVID-19 pandemic has affected
America, the slower-moving but impactful effects of political polarization have fractured trust in
and closeness between each other (Ivengar et al., 2012) and have eroded trust in well-cemented
systems like higher education – which has led to impacts on policy, such reducing appropriations
to public institutions (Dar & Lee, 2014; Taylor et al., 2020). Often, political polarization has less
to do with an attachment to ideological or policy preference and more of a social identity
(Grossman & Hopkins, 2016) that encourages individuals to curate closed social networks, attack
“out-group” members (Chen & Rohla, 2018; Iyengar et al., 2012; 2019), and be more compliant
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 6
or trusting with power structures associated with “in-group” leaders (Krupenkin, 2020).
Although both parties have experienced polarization, the cultivation of partisan identity has
seemingly been stronger for the Republican Party (Hare & Poole, 2014), which has likely been
strengthened by cultural homogenization as white people have recently more strongly identified
with and flowed into the Republican Party (Zingher, 2018) and the rise of the identity of being a
Trump supporter (Doherty et al., 2017; Jones, 2020). For reference, the Republican Party is
predominantly White (81%), over 50 years old (56%), and 70% do not have a college degree; the
Democrats are more diverse but mostly White (59%), are slightly younger as only 50% are 50
years old or over, and consist of more college-educated individuals (41%; Gramlich, 2020).
Polling and emergent studies illustrate a link between divergent political identities and
compliance related to COVID-19. Compared to those who identify (or lean) as Democratic,
individuals who identify (or lean) Republican have been less likely to view COVID-19 as a
major health threat (Tyson, 2020), wear masks in public (Kramer, 2020), and have shown a
lower intention to get vaccinated (Funk & Tyson, 2020). The differences in views of COVID-19
have also correlated to other policy preferences such as reopening K-12 schools (Menasce-
Horowitz, 2020). For higher education, Collier et al. (2020b) found that state-level Republican
leadership was correlated with an increased chance of institutions choosing in-person instruction
for Fall 2020 – a wildly popular partisan decision with the “in-group” base as 74% of polled
Republicans indicated that universities/colleges providing in-person learning were making the
“right decision to bring students back to campus” in Fall 2020; just 29% of Democrats felt this
way (Parker et al., 2020).
Partisanship and Higher Education Finances & Decision-making. Before tensions
surrounding COVID-19, political partisanship had been linked with eroding trust in higher
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 7
education (Doherty et al., 2017), behaviors related to adjusting state-based financial support (Dar
& Lee, 2014; Taylor et al., 2020), and responses to emergent higher education policies, such as
tuition-free college (Collier et al., 2019). State (and local) governments have generally been bent
to the will of the national party platforms (Hopkins, 2018). Furthermore, partisanship forcefully
guides policymaking and sets the terms for subsequent victories (Miller & Morphew, 2017).
Given those parameters, we should expect to see partisanship be linked with institutional support
(Dar & Lee, 2014; Taylor, et al., 2020) – and from the perspective of our study, linked with
institutional decision-making based on prior research (Collier et al., 2020b).
Nearly all higher education institutions are reliant on government support for financial
stability (through e.g. tax breaks, grants, appropriations, and other political favors). Generally,
governmental resource dependency is examined through public institutions (Fowles, 2014;
Tollefson, 2009) – however, examples can also be found for private institutions (see – Looney &
Lee, 2019; Muller, 1987). As such, much of the literature in higher education that is focused on
politics and partisanship discuss the effects of state leadership and economic factors on state
support for public institutions (Tandberg, 2010; McLendon et al., 2009; Okunade, 2014). During
recessions, public colleges expect to see large declines in state appropriations as the sector acts
as a balancing wheel for strained budgets (Delaney & Doyle, 2011; Hovey, 1999). Institutions
typically respond by increasing enrollment, raising tuition, and/or cutting services Colleges rely
on non-resident students who are willing to pay higher tuition prices to plug budget deficits
(Jaquette & Curs, 2015), but this option was limited in 2020 due to travel restrictions and
regulations making international student enrollment more difficult (Whitford, 2020).
More recent work has examined the relationship between higher education funding and
partisan control of state legislatures and governorships. Republican governors and legislatures
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 8
have typically been associated with lower appropriations for public higher education institutions
(Dar & Lee, 2014), unless White students were overrepresented (Taylor et al., 2020); therefore,
suggesting benefits exist for institutions that enroll students more closely related to the politically
in charge “in-group.” The links of partisanship on the financial benefits that public institutions
receive are just recently being fleshed out and more work is to be done. While Collier et al.
(2020b) found linkages between state legislature political power and 4-year private institutions
being more likely to choose in-person instruction, the effects of partisanship and benefits private
institutions receive from state and local governments as related to institutional decision-making
remains limited, despite private schools also relying on various governmental benefits for
increased stability (Muller, 1987). Moreover, the relationship between partisanship or
sociopolitical identity and decision-making at higher education institutions is currently limited.
As polarization is generally driven by social identity and not necessarily bounded by
policy preferences (Iyengar et al., 2012; 2019) and given that higher education institutions are
dependent on the government for financial and non-financial benefits (Fowles, 2014), it would
make sense that institutional decision-making would more closely align with currently held
power structures. That is, decision-makers and administrators may exhibit behaviors more
aligned with the “in-group” to maintain benefits and avoid negative sentiments associated with
being part of the “out-group” (see Billig & Tjfel, 1973 for more about “in” and “out” group
dynamics) rather than selecting behaviors based non-sociopolitical policy preferences.
Potentially, decision-makers are consistently weighing the short- and long-term risks of opposing
current power structures as related to the various dependencies that exist or are already partisans
themselves and want to avoid the scrutiny of being part of the “out-group.” Given that the
Democratic and Republican parties hold partisan preferences around both the state of higher
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 9
education in the United States (Parker, 2019) and policy responses to COVID-19 and its
mitigation (Corder, et al., 2020; Holman et al., 2020), we might expect differential responses to
COVID-19 based on the political characteristics of an institution’s locale (see Collier et al.,
2020b; Felson & Adamczyk, 2021).
Conceptual Framework & Hypothesized Model
Our model takes cues from and expands upon Collier et al. (2020b), the first study using
the C2i dataset to suggest that alignment with state political power structures and risks associated
with resource dependency (more so than the state or local severity of COVID-19) was linked
with institutional decisions to reopen in the Fall of 2020 as predominately in-person. Figure 1
illustrates our hypothesized model for the relationship between sociopolitical features, COVID-
19, state revenue changes, and in-person reopening. As highlighted by Figure 1, our model
includes three latent constructs: State Sociopolitical Features, County Sociopolitical Features,
Pandemic Severity and two observed variables State Revenue Changes and In-Person Instruction.
In the analysis section, we detail how these latent variables were generated, but here it is
important to understand these variables captured unmeasured attributes by essentially grouping
observed variables.
We generated a direct link from State Sociopolitical Features to County Sociopolitical
Features, where institutions’ campuses were geographically located. An influence from State- to
County Sociopolitical Features is theoretically sound given that – due to polarization and
cultural identification – local politics have seemingly been bent to the will of party-line goals
(see Hopkins, 2018). Additionally, as counties look to state governments for guidance on policy
and finances (Gold & Ritchie, 1992), this pathway makes particular sense when considering that
county-level governmental factors alone are not always sufficient in influencing economic
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 10
outcomes (Pink-Harper, 2018). Specifically related to predictors of County Sociopolitical
Features, prior evidence shows that state leadership can adjust policy (and rhetoric) to give
certain political outcomes a better chance (Ansolabehere & Snyder, 2006; Cahan, 2019) and
engineer social outcomes, such as college attainment (see Perna & Finnery, 2014).
Next, a direct pathway was generated from both State- and County Sociopolitical
Features to Pandemic Severity based upon emergent studies illustrating existing differences
between the dominant political parties (Republican and Democrat) and their framing of, and
policy responses to, COVID-19 (Dunn, 2020; Hartney & Finger, 2020; Holman, et al., 2020);
these differences have also manifested in distinct views between citizens who identify as
members of each party (Kramer, 2020; Parker, et al., 2020; Tyson, 2020). Given that polarization
is more of a social identity (Iyengar et al., 2012; 2019), the sociopolitical latent variables
predicted the percentage of those without a 4-year degree. Individuals with less than 4-year
degrees often hold distinct views from individuals with higher educational attainment in policy
preferences concerning the COVID pandemic – for example, a higher percentage of those with
4-year degrees feel COVID-19 is a “significant” crisis (75% v. 66% for those with some college
and 61% for those with H.S. degrees). Emergent research has linked these observed state
attributes with state-level COVID cases (Chambless, 2020; Frey, 2020).
Additionally, pathways were connected to State Revenue Changes from State
Sociopolitical Features, County Sociopolitical Features, and Pandemic Severity. State revenue
from taxes was tabulated from March to May 2020, as the pandemic began accelerating, and
compared against the same timeframe for 2019 to provide a change statistic (National Public
Radio, 2020). Emergent research has identified lockdowns were influenced by politics – in that
Democratic governors were three times more likely than Republican governors to impose a
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 11
lockdown (Tellis et al., 2020) and were more likely to engage a lockdown sooner (Corder, et al.,
2020). Given the timelines, due to locations of initial outbreaks in the U.S., “Blue” states won by
Hillary Clinton in 2016 experienced per-capita higher infection rates in March through June, than
did “Red” states won by Donald Trump– however, towards in June the trend converged as both
“Blue” and “Red” states’ infection rates are aligned (Barrow et al., 2020). As such the
combination of Pandemic Severity and State Sociopolitical Features likely influenced early State
Revenue Changes – again, justifying our linkage. Furthermore, given the lack of a strong federal
response to COVID-19, policy decisions were predominantly left to states; state-level responses
were uneven and sometimes disconnected with COVID-19 (Kettl, 2020).
Finally, guided by Collier et al. (2020b) each of the latent constructs in this model was
directly linked with the outcome of in-person instruction. In short, that study found that state
political control. Although per capita COVID cases were not generally correlated to the outcome,
county cases even less so than state cases; our study is different in that we captured COIVD
cases per capita at the time individual institutions made their last recorded decision – whereas the
two analyses in the descriptive study used cumulative COVID cases on June 15 and on August 5
regardless of when institutions made their decisions (Collier et al., 2020b). One aspect the
Collier et al. study did not examine was County Sociopolitical Features on the outcome. Given
both that the CDC suggests collaboration between local and state health officials (Centers for
Disease Control and Prevention, 2020) and that institutional leaders have previously asserted
they were following the guidelines of local political leaders and local government health officials
(Bauer-Wolf, 2020), a direct connection between county sociopolitical features and our main
outcome is appropriate.
[Figure 1 about here.]
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 12
Methodology
Sample
The College Crisis Initiative (C2i) database houses over 2,900 observed institutions. We
eliminated private for-profit and certificate-granting institutions and then placed institutions into
four sectors (2-year public, 4-year public, 2-year private, and 4-year private sectors) based on
2018 Carnegie classifications and sector control variables from the Integrated Postsecondary
Education Data System (IPEDS; National Center for Education Statistics, 2020). This resulted in
a final sample of N=2,469 institutions, of which n=940 (38%) were 2-year public, n=902 (37%)
were 4-year private, n=516 (20%) were 4-year public, and n=111 (4%) were 2-year private
institutions.
Variables and Data Sources
Our outcome of interest, from C2i data, is whether a college planned to have the Fall
2020 term “primarily” or “fully” in person as of September 9, 2020. This represents a college’s
final recorded decision for that date, meaning that institutions like the University of North
Carolina at Chapel Hill that abandoned in-person instruction in late August are counted as not
being in-person. We address concerns surrounding this gatekept variable in the limitations
section. In total 24% (n=596) institutions engaged in-person instruction - by institutional sector,
34% of 4-year private institutions engaged in in-person learning followed by 4-year public
(27%), 2-year private (22%), and 2-year public (13%) institutions.
We joined the C2i dataset with external data sources to consider multiple variables that
prior research suggests may correlate with institutional decision-making (see Castiello &
Whatley, 2020; Collier et al., 2020b). Daily COVID-19 cases by state were obtained from The
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 13
New York Times (2020b) and daily COVID-19 cases by county from USAFacts (2020).
1
The
COVID-19 data in our analyses reflect case rates at the date each institution made their last
known instruction decision from March 1 through September 9, 2020, which was by the time
instruction was generally slated to begin at the institution. The earliest decisions were made on
March 5, 2020, and the latest on September 7, 2020. For state and county cases, our data sources
provided a daily cumulative case count, and we calculated a 14-day moving average of new
cases per 100,000 residents at the time of the institution’s decision.
2
Next, we turned to the American Community Survey (ACS; U.S. Census Bureau, 2020)
to introduce sociopolitical state- and county-level attributes. We used the 2018 5-year estimates,
joining the percent of residents without educational attainment of a bachelor’s degree or higher.
We also tested other variables that we did not report in the main analysis – we provide more
details in the “alternative models” section.
We used data from the National Conference of State Legislatures (NCSL; 2020) to
identify the states in which both the legislature and governorship were under joint Republican
control – this was a binary outcome where 0=No, 1=Yes. For county-level political attributes, we
imported the share of votes for the Republican nominee for President (Donald Trump) in the
2016 presidential election (Tay, 2018). Using the share of votes for a given candidate in a
presidential election to proxy for a county’s political disposition is commonplace (see Frey,
1
In step with Collier et al. (2020), we chose USAFacts (2020) for county cases because the New York Times county
cases database combines all five counties into a single New York City-wide measurement.
2
To calculate the 14-day moving average of new COVID-19 cases, we first tabulated the new cases each day (i.e.,
the one-day change in cases) from the cumulative case count data. We accomplished this by subtracting the previous
day’s cumulative case count from the current day’s cumulative case count within each county/state. To get the 14-
day moving average of new cases, we then averaged the one-day change in cases for the current day and the 13
preceding days. We only generated an average value if there were 14 days of data available, so we do not have a 14-
day moving average until the 14th day a county/state reported COVID-19 cases. We generated a 14-day moving
average of new cases per 100,000 residents by multiplying the 14-day moving average of new cases in each
county/state by 100,000 and then dividing by the county’s/state’s total population.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 14
2020; Oberhauser et al., 2019). Given that Washington, DC has no governor or state legislature,
we have opted to drop DC-located institutions from our analyses (n=8), which has already been
factored into our N=2,469 final sample.
As a final component of our main model, we turned to NPR’s tabulation on state revenue
changes using data housed at the Urban Institute’s State and Local Finance Initiative. In August
2020, the NPR report compared the state tax revenue generated in March through May 2020 (a
three-month average) and compared the outcome against the same three months in 2019. Across
the U.S. the average change in state tax revenue was -29%. During this timeframe, three states
were still experiencing growth: North Dakota (+8%), Nevada (+5%), and South Dakota (+1%);
however, the rest were in decline with California (-42%), Alaska (-45%), and Oregon (-53%)
experiencing the largest declines (National Public Radio, 2020). Finally, we mean-centered this
variable. The NPR calculations were missing for the State of New Mexico (n=30). Instead of
dropping schools from this state, we imputed the data after mean centering, setting these missing
values at the average revenue change (so that in analyses, NM-located schools have a value of 0
on the mean-centered variable).
Based on our hypothesized model structure, we also joined included several variables
from the Integrated Postsecondary Education Data System (IPEDS). The IPEDS data allowed us
to examine the influence of institutional characteristics, via the importance of two latent
constructs which are shown only in appendices because they weakened the SEM. Percent out-of-
state enrollment represents the number of first-time undergraduate students who reported a state
of residence other than the state in which the institution is located divided by total first-time
undergraduate students. Percent room capacity represents an institution’s dormitory capacity
divided by total undergraduate enrollment. We also generated dummy variables to indicate
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 15
whether schools were in a Power 5 sports conference and whether their selectivity was deemed at
least Very Competitive in the Barron’s ranking (National Center for Education Statistics, 2020)
Structural Equation Modeling
Structural Equation Modeling (SEM) directionally examines the direct effect of one
variable on another and through this directionality, the technique also captures any indirect
influences from one variable passing through another (Klem, 2000). Furthermore, SEM, which
computationally builds on a set of regressions, can use several observed variables to jointly
measure a latent construct (of which some characteristics are not measured), and the overall
analyses relate the relationship of this latent construct (not its measures) to the other parts of the
model. In the past, researchers have used SEM to identify factors influencing first-year college
persistence (Collier et al., 2020a), student growth and development (Wofford, 2020), and
administrative decisions surrounding expenditures or adoption of new technology (El-Masri &
Tarhini, 2017). Relatively few pieces have attempted to integrate institutions’ contextual
sociopolitical factors into SEM models examining higher education outcomes. Some have
explored individual-level attributes like veteran status or partisan political preferences (Gonzalez
& Elliott, 2016) or social orientations toward people, community building, and leadership (Harris
et al., 2016). But pieces leveraging SEM to explore institutional-level decisions focus more on
institutional characteristics and approaches (Manzoor et al., 2020), or faculty-specific
orientations toward pedagogy (Masserini et al., 2018) rather than sociopolitical ones. It is likely
that our analysis adds novel nuance in its consideration of state and county features as well as
being the first that we are aware of to use SEM to examine the outcome of in-person instruction
during the COVID-19 pandemic.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 16
To use SEM, we must remain compliant with several core assumptions highlighted by
Kline (2012). First, we must maintain temporal sequencing, meaning that the main outcome
variable must occur after all observed information – in this case, the decision to be in-person
occurred after each of the observed variables (often also called measured variables) in the model.
Temporal sequencing in SEM limits the possible directions of the relationships examined. Our
model remains in step with temporal sequencing given that the main outcome of choosing in-
person instruction occurred after all other factors in the model, State- and County Sociopolitical
Features were already set before institutions made decisions and decisions were likely in
response to Pandemic Severity and State Revenue Changes brought about by the pandemic.
Second, observed variables included in the model must be correlated with the main outcome. As
Table 1 illustrates, all observed variables in the models are significantly correlated with the main
outcome of in-person instruction.
[Table 1 about here.]
Third, the proper statistical approaches for an SEM analysis must be employed, as there
are multiple estimators – for example, the commonly used Maximum Likelihood Approach
(ML). As our main outcome was binary, 0=not in-person, 1=in-person; we used a weighted-least
square means and variance-adjusted approach (WLSMV). The same approach can be found in
prior research (Bowman, et al., 2019; Collier et al., 2020a). For models with a binary main
outcome, a WLSMV approach is more appropriate than an ML estimator, as the WLSMV has
been shown to produce more accurate factor loadings, interfactor correlations, and structural
coefficient estimates (DiStefano & Morgan, 2014; Li, 2016). To test our models, we used the
freeware JASP, which includes the Lavaan module also found in R. Although SEM studies have
recently gained traction, an ongoing debate remains over how to determine what measures
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 17
should be used to statistically accept these models. For guidance among measures, we leaned on
a combination of prior empirical studies published in strong higher education journals (see
Bowman et al., 2019; Collier, et al., 2020a) and methodological works (Hu & Bentler, 1998; Xia
& Yang, 2019).
Based on best practices for the specific analytic structure we employ, we rely on and
report the Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) for the goodness of fit,
and the Root Mean Square Error of Approximation (RMSEA) and the Standardized Root Square
Residual (SRMR) for the badness of fit. The aforementioned methodologists suggest that SEM
tested models are statistically acceptable if CFI≥.95, TLI≥.95, RMSEA≤.06, and SRMR≤.08.
There is also an optimal fit range where CFI≥.98, TLI≥.98, RMSEA≤.03, and SRMR≤.07 (Xia &
Yang, 2019). The overall model fit statistics are acceptable, as are the fit for 4-year private and
2-year public institutions. However, the fit for the 4-year public institutions is within the optimal
range - see Table 2 for model fit statistics. We did not rely on, nor report, the Chi-Square statistic
because this measurement is an inadequate measurement to assess SEM analyses with larger
sample sizes (Kenny & McCoach, 2003), such as ours.
[Table 2 about here.]
Finally, the reported values are robust standardized coefficients (for example, r=.xx), which
report the relative size of the influence on the mean of the tested variable. Standardized
coefficients allow for comparisons of variable impact by changing all of the variables to a mean
of 0 and then measuring the impact of an increase of one standard deviation in the predictor
variable on the tested variable (Kwan & Chan, 2011).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 18
Development of Latent Constructs
In SEM, variables can be observed or latent – a latent construct is an unobserved
variable, but information about which can be understood from observed variables. To generate
latent constructs for this study we conducted an exploratory factor analysis (EFA). EFAs are
employed when researchers do not have a specific hypothesis surrounding the nature of the
underlying factor structure and want to identify clusters of variables that could be grouped, and
isolate constructs from observed variables (Yong & Pearce, 2013). We generated an EFA and
identified three latent constructs. We labeled the first construct State Sociopolitical Features
which predicted the observed variables of Republican governmental control and the percentage
of adults 25 and older without a Bachelor’s degree or higher. The second we labeled County
Sociopolitical Features which predicted the share of Republican vote in the 2016 presidential
election and percentage of adults 25 and older without a Bachelor’s degree or higher.
3
The third
construct variable was Pandemic Severity which predicted the observed state and county 14-day
average of COVID-19 cases per capita at 100k persons at the time of institutions’ last reported
decision. Observed variables’ R2 values are reported in Table 3; the R2 reports the fraction of
variance explained by each variable for the outcome variable.
[Table 3 about here.]
3
Conversations with colleagues have resulted in debate on whether the observed variable of percentage with
bachelor’s degree could be included in both State and County Sociopolitical Features – as state features predict
county features in our model. Given that latent variables measure the “unobserved” and predict observed variables,
it is conceivable that two distinct latent constructs at two geographic levels both predict (in addition to unique
measured variables) one variable in common. In consideration of the fact that reviewers and readers may hold
similar concerns, however, we have generated alternative models where County Sociopolitical Features is replaced
with only an observed variable of the share voters who voted for the GOP candidate in 2016 (Table A1). Generally,
the trends associated with significance remain the same in the underlying structure, as do the magnitude of the
coefficients (although the point estimates are different as would be expected). Similar outcomes are also found for
the main outcome of in-person instruction – the only difference is that in the 4-year public institutions County
Sociopolitical Features is significant (r=.11, p=.03), whereas in our reported model the influence was not (r=.12,
p=.20). The model statistics are no stronger for this variant approach; we retain the latent construct based on our
understanding of how theory predicts the interaction of state, county, and university forces.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 19
Limitations
As noted, our data on mode of instruction and pandemic severity are uniformly from
September 9th, 2020, and do not fully represent decisions made at the time classes first started
(i.e. a range of dates across August and September that are institution-specific). At the time of
analysis, the C2i dataset did not include Fall 2020 start dates for all institutions – although it has
become a focus for C2i, at the time of writing this manuscript they have start dates for less than
half of the institutions in the dataset. To be sure that our analyses generally captured the decision
at the start of classes and not, a result of a mass pivoting forced by high COVID cases on a
campus that attempted to start in-person classes, we tabulated how many institutions had made
their last on-record decisions from August 17 to September 9, 2020. We chose August 17 as the
University of North Carolina moved from being in-person to online, which is arguably one of the
more publicized pivots of the Fall semester. During the August 17 to September 9 timeframe,
n=58 (6%) 2-year public, n=36 (4%) 4-year private, n=22 (4%) 4-year public, and n=7 (6%) 2-
year private institutions made any operational decisions. Given that just 5% of the total sample
(and no more than 6% of institutional sector groups) made decisions in this later timeframe, we
believe our findings generally represent influences on the initial plans made in advance of Fall
reopening. For fidelity, we checked our main model with these institutions removed – the
goodness and badness of fit statistics remained the same as did the significance of variables on
the main outcome (although some coefficients experienced minor changes in point estimate). We
also generated models for the sectors examined and found no difference in fit statistics or the
trends of significance. In the future, using data on the decision by the start of classes – and
associated county and state date-dependent COVID data – for institutions rather than September
9th should improvement on our work.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 20
Next, given the prior understanding of the importance of financial pressures to
administrator decision-making (see Delaney & Doyle, 2011), we strove to include additional data
on within-2020 budget changes but were unable to do so. We looked at New America’s efforts in
tracking higher education budget declines (Nguyen, et al., 2020) and debated including these
data into our model. We decided against it, as the information was updated in October well after
institutions made decisions (i.e. it reflected pre-pandemic budgetary realities) and the report only
listed information for 44 states. We decided to exclude this information as we believed early
state revenue declines for 49 of 50 states might be a better indicator, given the timeframe when
decisions were being made, even though it is less specific than higher education budgets
(National Public Radio, 2020). By the time future researchers can retest and expand upon our
models, there will likely be more information on state revenue declines, and, specifically budget
cuts to higher education – of which we encourage future inclusion.
Finally, one missing key attribute in our models may be information regarding individual
institutions’ governing boards (see Morgan, et al., 2020). We wanted to include institutional
governing board make up; however, given the diversity of these boards, lack of individual board
members’ attributes, and how board members are appointed we could not include the variable in
this analysis. These boards may have been a mechanism through which local politics and
identities are formalized in campus contexts, constraining the choice set administrators had even
on rapid-turnaround decisions on which the boards had no observable influence. Our work did
not examine this question, but later qualitative work could fruitfully explore this possibility.
Likely, future research expanding these models may find value in including governing board
attributes – when such a dataset exists.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 21
Findings and Discussion
In this section, we first focus on the overall model – in which all institutions in our
sample are included in the model structure. Initially we highlight and discuss the findings in the
underlying structure (i.e., the relationships between the latent constructs and observed variables
leading up to the main outcome), and then focus attention on what influenced the main outcome
of reopening at least primarily in-person. Then, we detail and discuss the findings for three main
institutional sectors: 4-year public, 4-year private, and 2-year public institutions. See Figure 2 for
a visual representation of significant direct pathways and Table 4 for all direct, indirect, and total
path coefficients. Please also see Figure A1, in the appendix, for the technical layout of the
underlying structure.
[Figure 2 about here]
[Table 4 about here]
Overall Model
Underlying Structure. Our main model has identified that State Sociopolitical Features
(r=.57) influence County Sociopolitical Features – aligning with research beyond higher
education suggesting similar outcomes (see Gold & Ritchie, 1992). As partisanship has advanced
and issues are now more strongly focused through party agendas (Mason, 2018), county-level
sociopolitical autonomy has likely been eroded, similarly as what occurred between the state and
federal levels – in that policy decisions may be more a referendum of political and cultural
identity and not ideology (Hopkins, 2018; Iyengar et al., 2019). We suggest the underlying
structure may be detailing the tightening of the party and cultural influences over local politics as
detailed by Hopkins (2018).
4
4
This outcome may explain why the exploratory model where we tested influences from county to state features
were demonstrably weaker and statistically unacceptable.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 22
Despite the latent state and county features constructs predicting similar observed
variables – Republican-leaning and lower-educational-attainment populations – their influences
on Pandemic Severity at the time of institutional decisions diverge. Overall, State Sociopolitical
Features (r=.40) had a positive influence on COVID per-capita cases, whereas, County
Sociopolitical Features had a negative influence (r=-.23). These divergent trends may signal two
interesting points. First, although State Sociopolitical Features influenced County Sociopolitical
Features, shoring up prior suppositions that local political autonomy is diminishing (Hopkins,
2018), at the time of institutional decisions, it is possible that decisions responding to Pandemic
Severity within counties hosting higher education institutions remained more autonomous from
wider political and cultural features.
The second piece may be that institutions in places with certain County Sociopolitical
Features made decisions earlier when COVID cases were not relatively high. Guided by Frey’s
(2020) trends, we generated a correlation matrix between our observed County Sociopolitical
Features and institutions’ selecting in-person instruction by June 1 (21% of institutions). To be
noted, n=527 (21%) of institutions in our sample made their last reported decision ‘early’: by
June 1; of these institutions n=484 (20% of the total sample; 91% of early deciders) chose in-
person instruction. We found positive correlations between an “early’ in-person decision and
county GOP vote-share (Pearson’s r=.16, p<.001) and percent of the county without a Bachelor’s
degree (Pearson’s r=.05, p=.01). This linkage suggests that the same power structures and
preferences associated with County Sociopolitical Features also encouraged institutions to make
decisions earlier (i.e. without information on the status of the pandemic closer to the schools’
opening date) which aligns with our framework of polarization and partisanship dominating
policy preferences (see Collier, et al., 2020b; Grossman & Hopkins, 2016; Taylor, et al., 2020).
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 23
On the other hand, since the pandemic did not uniformly sweep across the nation, it
remains possible that the policy preferences associated with county features were based upon
lower per capita COVID-19 cases at the time of the known decision. Again, we turned to the
observed variables predicted by the latent constructs. Conditional upon an ‘early’ decision, we
found a correlation between county share of vote for the GOP candidate and state COVID cases
per capita (Pearson’s r=.09, p=<.001) but not county cases per capita. We also found
relationships between the percent of county residents without a Bachelor’s degree and both state
COVID-19 cases per capita (Pearson’s r=.24, p=<.001) and county cases per capita (Pearson’s
r=.09, p=<.001). Given that positive correlations exist between the measured variables predicted
by the County Sociopolitical Features and the measured variables of Pandemic Severity
conditional upon this ‘early’ decision – we believe these outcomes reinforce narratives that
sociopolitical preferences likely superseded “the science.”
Finally, as related to early State Revenue Changes we found State Sociopolitical Features
(r=.38) were the strongest influence and positively correlated, meaning state features that predict
Republican leadership and below Bachelor’s education influenced lower tax revenue declines.
This outcome makes sense as states with Republican governors were generally less stringent than
Democrats on following suggested policies – such as resisting restricting mobility (Akovali &
Yilmaz, 2020) or delayed lockdowns (Corder, et al., 2020; Tellis et al., 2020) – mitigation tactics
which may lead to more severe state revenue declines. County Sociopolitical Features (r=.10)
were also impactful in the same direction as state features. However, the influence on revenue
changes was not only sociopolitically guided, as Pandemic Severity (r=-.05) negatively
influenced revenue – albeit to a substantially lower degree than sociopolitical features. In
combination these outcomes align with the general premise that sociopolitical features predicted
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 24
by these latent constructs did not view COVID-19 as a similar serious threat as those in the “out-
group” (see Tyson, 2020).
Main Outcome of In-Person Instruction. Table 4 reveals that three of the four variables
had a significant total (direct plus indirect) influence on the main outcome of primarily being in
person. The strongest influence on choosing in-person instruction for Fall 2020 was County
Sociopolitical Features (r=.13), followed by Pandemic Severity (r=-.10), then State
Sociopolitical Features (r=.09). Given how close the strength of influence from County
Sociopolitical Features and Pandemic Severity are, we are not willing to say for certain that one
factor was absolutely “more” influential than the other based on the relative size of the point
estimates. Although both levels of sociopolitical features significantly affected the outcome,
seemingly institutions were either more receptive or sensitive to county-level features (or were
more likely to work with local decision-makers). Still, given that county and state features
influenced the outcome in a positive direction, we again (see Collier et al., 2020b; Felson &
Adamczyk, 2021) have evidence of sociopolitical features encouraging behavior aligned with
preferences that minimized the severity of COVID-19 (Tyson, 2020). As a reminder, in-person
instruction was connected with an increase of roughly 6,500 new cases of COVID-19 per day
nationwide (Andersen et al., 2021).
Pandemic Severity was the only negative influence on the outcome – lining up with an
expectation that the more serious a state and county’s infection rates were, the less likely an
institution would be to resume in-person instruction and bring students back to campus. This
outcome contests prior descriptive research suggesting no clear relationship between COVID-19
cases at either the state or county level and institutional decisions to select in-person instruction
(Collier et al., 2020b; Felson & Adamczyk, 2021). However, we must remind readers that this
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 25
paper is different in that we used COVID case rates at the time of an institution’s last recorded
decision by September 9, 2020 to develop a latent construct; whereas, the former papers
generated individual state and county rates based on cumulative totals at specific timeframes.
Given our findings, we can suggest that in aggregate, institutional decision-makers were
considering and monitoring the severity of COVID-19 as they suggested they would (Bauer-
Wolf, 2020). However, this overall finding can mask important differences in how responses to
pandemic severity and other pressures differed by institutional characteristics.
Trends Across Institutional Sectors
Our main model answered the question of how the factors examined influenced the sum
of higher education to operate in-person. Yet, in understanding that sectors of higher education
may respond differently to distinct pressures, we also generated separate models for 4-year
public, 4-year private, and 2-year public institutions. It seemed particularly likely that we would
observe differences given the distinct funding structures and local community integration of the
institution types. In this section, we detail and discuss the patterns and contrasts across
institutional categories.
When examining these sectors separately, we found that Pandemic Severity only
significantly influenced 4-year private institutional decisions to engage in-person instruction (r=-
.09). The coefficients for 4-year public (r=-.07) and 2-year public (r=-.07) were similar to each
other and did not differ greatly in magnitude from 4-year private institutions, with all flowing in
a negative direction – as would be hoped for. Given that Pandemic Severity was not significant
for 4-year and 2-year public institutions, we interpret this pattern to suggest that public
institutional decision-makers were simply not particularly sensitive to Pandemic Severity at the
time of decision. As public institutions like the University of North Carolina have made widely-
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 26
publicized missteps related to in-person instruction (Flaherty, 2020; Treisman, 2020) and
nationwide delivery of in-person instruction has resulted in thousands of additional cases per day
(Andersen et al., 2020), these findings place decision-makers at public institutions in an
uncomplimentary light. While 4-year private institutions were more sensitive to Pandemic
Severity than their peers, which should be praised, sociopolitical features were generally stronger
influences on the main outcome for all types of schools.
First, State Sociopolitical Features significantly affected each sector examined, with the
largest influence found for 4-year public institutions (r=.17), followed by 4-year private (r=.11)
and 2-year public (r=.07) institutions. Given the deep body of research on 4-year public
institutions and their dependency on state governments (see Delaney & Doyle, 2011; Fowles,
2014) and the effects of polarization on appropriations to 4-year public institutions (Dar & Lee,
2014; Taylor et al., 2020), we expected that 4-year public institutions would be the most
sensitive to state-level sociopolitical features and likely thrive to remain close to the state-level
“in-group.” In combination, though, our findings indicate that State Sociopolitical Features sway
decisions made by institutions who are generally understudied within this focus. Given that each
of these sectors are in their own way sensitive to these features, each are likely making decisions
to avoid the risks of losing financial, political, or social support when being perceived as part of
the “out-group.”
Next, 4-year private (r=.28) and 2-year public (r=.23) institutions were also significantly
and substantially influenced by County Sociopolitical Features – while 4-year public institutions
(r=.12) were not. County-level influence was over double that of state-level influence for 4-year
private institutions and triple for 2-year public institutions; therefore, suggesting that these
institutions were more sensitive to more localized sociopolitical features. Two-year public
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 27
institutions have long been dependent on local resources and/or find themselves under local
control (Tollefson, 2009), which likely explains the outsized effect by County Sociopolitical
Features while exhibiting similar directions as State Sociopolitical Features due to the top-down
features of polarization (Hopkins, 2018).
It is less clear why 4-year private institutions would be influenced by both levels of
sociopolitical features. Yet, our findings align with prior studies suggesting that private
institutions are not quite as politically independent as many would believe (see Collier et al.,
2020b; McLendon et al., 2009; Okunade, 2004). What our models may be capturing is a
sensitivity of 4-year private institutions towards being part of the sociopolitical “in-group” –
essentially, the identification of risks with losing political favor and potentially the ability to
maintain or increase their market share via matching the cultural identity of their student-
constituent market. Furthermore, enrollment-dependent private institutions may recognize the
importance of taking cues from state and local policymakers as one part of maintaining an
environment (or appearance) of being competitive with public institutions. Prior research has
illustrated that for modest public investments, private institutions will accept additional
accountability and political oversight (Muller, 1987); and as such, may be influenced by
sociopolitical features to maintain a robust enrollment base and public investment (or political
favors; Richardson et al., 1999). Overall, our study affirms that private universities must be
considered political actors and illustrates the need for more studies examining how private
institutions respond to state and local sociopolitical features.
Alternative Specifications and Future Research
Based upon Collier et al. (2020b), we tested alternative models that included elements of
prestige and residential characteristics. EFA allowed us to generate a latent variable we called
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 28
Prestige Elements that predicted the following observed binary variables captured from IPEDS:
Baron’s Very Competitive or Higher, Power 5 Sport Conference, and Billion Dollar+
Endowment. We also generated a Residential Characteristics latent variable that predicted:
percentage of enrolled undergraduates that were out of state students and a percentage of room
capacity for the enrolled undergraduate population (National Center for Education Statistics,
2020) – see Figure A2 in the appendix. This model produced acceptable CFI=.95 fit, but
unacceptable TLI=.93, RMSEA=.10 and SRMR=.07.
A pared-down model eliminating the Prestige Elements latent variable produced a
slightly stronger model with acceptable CFL=.97, TLI=.96, SRMR=.05, and unacceptable
RMSEA=.08 – each of these fit statistics were weaker than our reported model fit. Moreover, the
model fits for the sector analyses were generally weaker. To be sure we were reporting the
strongest of the two models for the full sample (despite issues with the individual sectors), as
also found in Collier et al. (2020a) we used the SBDIFF.EXE program (Crawford & Henry,
2003) to conduct a DIFFTEST between this overall second model and our reported model,
finding the models were statistically different, as such we decided to report and focus our paper
on the stronger overall model. In these models, Residential Characteristics model was a non-
significant influence on the outcome of Fall in-person instruction. Potentially, models including
elements of prestige and residential characteristics may be more relevant for decisions related to
Spring 2021 reopening, as policy agendas would likely be more cemented, revenue declines and
associated cuts to institutions would be better realized, and decision-makers have gained
experience in dealing with COVID-19. We believe our main reported models and these models
should be re-tested and expanded upon as new information (e.g. campus mitigation strategies)
becomes available – as this is just the beginning of the conversation.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 29
Next, we tested our models with the 2020 Presidential Election results for county share of
votes (New York Times, 2020a) instead of the 2016 vote share. We conducted this sensitivity
analysis because an argument could be made that these data from November 2020 may better
reflect recent policy preferences and polarization, concerning decision-making during summer
2020, than an election 3.5 years earlier. These models mostly remained the same which is likely
due to an essential lack of change from 2016 to 2020 as the same share of the vote (47%) in these
counties went to President Trump. We opted to report the model with the 2016 data for two core
reasons. First, because it remained aligned with our temporal sequencing, as the 2020 election
occurred after institutions made decisions, whereas the 2016 vote share is a reasonable preceding
measure for summer 2020 institutional pressures (and thus aligned with the assumptions of
Kline, 2012). Second, and crucially, the pandemic itself inflamed partisan tensions, voter turnout,
and likely party identification; COVID-19 created pressures that both school reopening and voter
behavior were responding to. Although 2016 data may seem no longer recent, 2020 vote share
most certainly was changed by the phenomenon under examination, not least because the student
vote in counties with college towns was likely different in areas where institutions chose online-
only education. Given that the structure of both models’ outcomes remained essentially similar
and no point estimate differs across the models by more than 0.01, this second model serves as a
post hoc robustness check on our reported findings.
Conclusion
As the pandemic hit colleges and universities during the Spring 2020 semester and then
the novel coronavirus spread over the summer, institutions struggled to make decisions that
balanced revenue concerns and politics with public health guidance. This paper used a novel
dataset to examine the contributing factors to those decisions through structural equation
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 30
modeling. Expectedly, we found that both state and county sociopolitical features contributed to
campus decisions regarding their mode of instruction. That higher education made decisions tied
to county and state sociopolitical conditions are unsurprising (see Collier et al., 2020b). The
national COVID-19 response has been anything but uniform. States and localities have been
responsible for setting their response guidelines for their citizens. Community colleges, in
particular, rely heavily on county support; public universities rely on state support. We posit that
private non-profit institutions, which have to compete with public institutions, likely made
decisions to avoid being cast as part of the sociopolitical “out-group” to maintain political favors
and retain a cultural attachment to their enrollment pool as to not lose enrollment to local or
regional competitors. Qualitative work on the decision-making processes of each type of
institution is sorely needed to better understand the unique characteristics of the state/county-to-
private institution relationship.
Unlike prior research (Collier et al., 2020b; Felson & Adamczyk, 2021; Marsicano et al,
2020), we also found that Pandemic Severity influenced the decision, though only when
considering the aggregate of the sample or private institutions. Unfortunately, public institutions
were not necessarily sensitive to the severity of the pandemic when gauging whether to provide
in-person instruction. Further, although 4-year private institutions were sensitive to Pandemic
Severity, State Sociopolitical Features mattered slightly more, and County Sociopolitical
Features were nearly three times stronger on the main outcome. When considering these sector
influences, it becomes obvious that sociopolitical features and likely intentions of being part of
the “in-group” associated with these features were simply more important factors when
considering an in-person reopening.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 31
Given that opening campuses for in-person instruction resulted in a daily increase of
6,500 COVID-19 cases across the country (Andersen, et al., 2021); polarization, political
identification, and institutions feeling compelled to operate as politized entities (to stay close to
the “in-group”) likely made the severity of the pandemic worse. Although our study is stronger,
these outcomes confirm the general findings of Collier et al. (2020b), that decision-makers did
not consider pandemic severity much in making choices about in-person instruction. In the likely
case that future studies confirm that appraisal, what is the sociopolitical fallout for institutions
when policymakers, partisans, and the public come to better understand higher education’s role
in the pandemic? Given that in-person instruction was heavily favored by Republicans but that
higher education is a consistent target for partisans and that trust in higher education has
experienced erosions, institutions’ behavior during the pandemic may have implications for
appropriations, regulations, and other policies in the long run.
Further research is necessary to understand under what conditions institutions took into
account public health information, whether there are systematic moderating or mediating factors,
and whether the inclusion of structural elements we omitted yet further alters the picture of how
important pandemic severity was to mode-of-instruction decisions. As institutions have been
making instructional decisions for Spring 2021, the nation recently hit its highest COVID-19 per-
capita case and death counts and a new, much more infectious and possibly deadlier mutation of
COVID-19 (U.K. variant; Hornby et al., 2021) has been detected in college students (Bowden,
2021). Given the severity of the pandemic, perhaps it is time for state and local infection rates to
become the primary determinant of future plans – but given the continued state of polarization
and policy preferences in response to the pandemic, we do not expect that will be the case.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 32
References
Akovali, U., & Yilmaz, K. (2020, December 10). Polarised pandemic response to COVID-19
connectedness across US states. VoxEU CEPR Policy Portal. Retrieved from
https://voxeu.org/article/polarised-pandemic-response-and-covid-19-connectedness-
across-us-states
Andersen, M. S., Bento, A. I., Basu, A., Marsicano, C., & Simon, K. (2021). College openings,
mobility, and the incidence of covid-19 cases (Working Paper). medRxiv.
https://doi.org/10.1101/2020.09.22.20196048
Ansolabehere, S., & Snyder Jr., J. M. (2006). Party control of state government and distribution
of public expenditures. The Scandinavian Journal of Economics, 108(4), 547-569. doi:
10.1111/j.1467-9442.2006.00470.x
Barrow, B., Vineys, K., & Kastanis, A. (2020, June 30). Coronavirus’ spread in GOP territory,
explained in 6 charts. Associated Press News. Retrieved from
https://apnews.com/article/7aa2fcf7955333834e01a7f9217c77d2
Bauer-Wolf, J. (2020, August 7). College say they’ll follow public health advice. What if they
don’t? Education Dive. Retrieved from https://www.educationdive.com/news/colleges-
say-theyll-follow-public-health-advice-what-if-they-dont/583139/
Billig, M, & Tajfel, H. (1973). Social categorization and similarity in intergroup behavior.
European Journal of Social Psychology, 3(1), 27-52. doi:10.1002/ejsp.2420030103
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 33
Bowden, J. (2021, Jan 24). University of Michigan halts athletic events after new COVID-19
strain detected. The Hill. https://thehill.com/homenews/state-watch/535587-university-of-
michigan-halts-athletic-events-after-new-covid-19-strain
Bowman, N. A., Miller, A., Woosley, S., Maxwell, N. P., & Kolze, M. J. (2019). Understanding
the link between noncognitive attributes and college retention. Research in Higher
Education, 60, 135-152. https://doi.org/10.1007/s11162-018-9508-0
Cahan, D. (2019). Electoral cycles in government employment: Evidence from U.S.
gubernatorial elections. European Economic Review, 111, 122-138.
https://doi.org/10.1016/j.euroecorev.2018.09.007
Castiello, S. & Whatley, M. (2021) Balancing safety and financial health: Non-resident student
enrollment and reopening plans at US higher education institutions amid the COVID-19
pandemic (Working Paper). Paper to be presented at the 2021 Annual Conference of the
Comparative and International Education Society.
Centers for Disease Control and Prevention. (2020, October 8). Coronavirus disease 2019
(COVID-19): Considerations for institutions of higher education. Washington, DC.
https://www.cdc.gov/coronavirus/2019-ncov/community/colleges-
universities/considerations.html
Chambless, L. (2020). Why do per capita COVID-19 case rates differ between U.S. states?
(Working Paper). medRxiv. doi:10.1101/2020.10.16.20213892
Chen, M. K., & Rohla, R. (2018). The effect of partisanship and political advertising on close
family ties. Science, 360(6392), 1020-1024. doi:10.1126/science.aaq1433
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 34
Chronicle of Higher Education, The. (2020, October 1). Here’s our list of colleges’ reopening
models. Retrieved from https://www.chronicle.com/article/heres-a-list-of-colleges-plans-
for-reopening-in-the-fall/
Crawford, J. R., & Henry, J. D. (2003). The depression anxiety stress scales: Normative data and
latent structure in a large non-clinical sample. British Journal of Clinical Psychology, 42,
111-131. doi: 10.1348/014466503321903544
Collier, D. A., Fitzpatrick, D., Brehm, C., Hearit, K., & Beach, A. (2020a). Structuring first-year
retention at a regional public institution: Validating and refining the structure of
Bowman’s SEM analysis. Research in Higher Education, OnlineFIRST.
doi:10.1007/s11162-020-09612-w
Collier, D. A., Fitzpatrick, D., Snideman, S., & Marsicano, C. (2020b). “What’d we miss?”: An
initial analysis of politics, demographics, and COVID-19 rates in colleges’ resumption of
instructional practices for Fall 2020 (Working Paper). American Political Science
Association, Washington, DC. doi:10.33774/apsa-2020-4c0f2
Collier, D. A., Mishra, S., Houston, D. A., Hensley, B. O., & Hartlep, N. D. (2019). Americans
‘support’ the idea of tuition-free college: An exploration of sentiment and political
identity signals otherwise. Journal of Further and Higher Education, 43(3), 347-362.
doi:10.1080/0309877X.2017.1361516
Corder, J. K., Mingus, M. S., & Blinova, D. (2020). Factors motivating the timing of COVID-19
shelter in place orders by U.S. Governors. Policy Design and Practice, OnlineFIRST.
doi:10.1080/25741292.2020.1825156
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 35
Dar, L., Lee, D.W. (2014). Partisanship, political polarization, and state higher education
budgets. The Journal of Higher Education, 85(4), 469-498.
doi:10.1080/00221546.2014.11777337
Delaney, J. A., & Doyle, W. R. (2011). State spending on higher education: Testing the balance
wheel over time. Journal of Education Finance, 36(4), 343-368.
https://www.jstor.org/stable/23018116
Denny, T. N., Andrews, L., Bonsignori, M., Cavanaugh, K., Datto, M. B., Deckard, A., ... &
Wolfe, C. R. (2020). Implementation of a pooled surveillance testing program for
asymptomatic SARS-CoV-2 infections on a college campus—Duke University, Durham,
North Carolina, August 2–October 11, 2020. Morbidity and Mortality Weekly
Report, 69(46), 1743. https://doi.org/10.15585/mmwr.mm6946e1
DiStefano, C., & Morgan, G. B. (2014). A comparison of diagonal weighted least squares robust
estimation techniques for ordinal data. Structural Equation Modeling: A
Multidisciplinary Journal, 21(3), 425-438. doi:10.1080/10705511.2014.915373
Doherty, C., Kiley, J., & Johnson, B. (2017). Sharp partisan divisions in views of national
institutions: Republicans increasingly say colleges have negative impact on U.S.
Washington, DC: Pew Research Center. Retrieved from
https://www.pewresearch.org/politics/wp-content/uploads/sites/4/2017/07/07-10-17-
Institutions-release.pdf
Dunn, A. (2020, Oct 21). Only 24% of Trump supporters view the coronavirus outbreak as a
‘very important’ voting issue. Washington, DC: Pew Research Center. Retrieved from
https://www.pewresearch.org/fact-tank/2020/10/21/only-24-of-trump-supporters-view-
the-coronavirus-outbreak-as-a-very-important-voting-issue/
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 36
El-Masri, M. & Tarhini, A. (2017). Factors affecting the adoption of e-learning systems in Qatar
and USA: Extending the Unified Theory of Acceptance and Use of Technology 2
(UTAUT2). Educational Technology Research and Development, 65(3), 743-763.
doi:10.1007/s11205-018-1927-y
Felson, J., & Adamczyk, A. (2021). Online or in person? Examining college decisions to reopen
during the COVID-19 pandemic in fall 2020. Socius: Sociological Research for a
Dynamic World, Online. doi:10.1177/2378023120988203
Flaherty, C. (2020, August 18). Shutdown, repeat. Inside Higher Ed. Retrieved from
https://www.insidehighered.com/news/2020/08/18/unc-chapel-hill-sends-students-home-
and-turns-remote-instruction
Fowles, J. (2014). Funding and focus: Resource dependence in public higher education.
Research in Higher Education, 55, 272-287. doi:10.1007/s11162-013-9311-x
Frey, W. H. (2020, October 8). As Election Day nears, COVID-19 spreads further into red
America. Washington, DC: Brookings Institute. Retrieved from
https://www.brookings.edu/blog/the-avenue/2020/10/08/as-election-day-nears-covid-19-
spreads-further-into-red-america/
Funk, C., & Tyson, A. (2020). Intent to get a COVID-19 vaccine rises to 60% as confidence in
research and development process increases. Washington, DC: Pew Research Center.
Retrieved from https://www.pewresearch.org/science/2020/12/03/intent-to-get-a-covid-
19-vaccine-rises-to-60-as-confidence-in-research-and-development-process-increases/
Horby, P., Huntley, C., Davies, N…. Semple, C. (2021). NERVTAG note on B.1.1.7 severity
(working paper). SAGE. Retrieved from
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 37
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/955239/NERVTAG_paper_on_variant_of_concern__VOC__B.1.1.7.pdf
Gold, S. D., & Ritchie, S. (1993). State policies affecting cities and counties in 1992. Public
Budgeting and Finance, 13(1), 3-18. doi:10.1111/1540-5850.00961
Gonzalez, C. A., & Elliott, M. (2016). Faculty attitudes and behaviors toward student veterans.
Journal of Postsecondary Education and Disability, 29(1), 35-46.
Gramlich, J. (2020, October 26). What the 2020 electorate looks like by party, race and ethnicity,
age, education and religion. Washington, DC: Pew Research Center. Retrieved from
https://www.pewresearch.org/fact-tank/2020/10/26/what-the-2020-electorate-looks-like-
by-party-race-and-ethnicity-age-education-and-religion/
Grossman, M., & Hopkins, D. A. (2016). Asymmetric politics: Ideological Republicans and
group interest Democrats. New York, NY: Oxford University Press.
Hare, C., & Poole, K. T. (2014). The polarization of contemporary American politics. Polity,
46(3), 411-429. doi:0032-3497/14
Harris, K., Hinds, L., Manansingh, S., Rubino, M., & Morote, E. S. (2016). What type of
leadership in higher education promotes job satisfaction and increases retention? Journal
for Leadership and Instruction, 2016(Spring), 27-32.
Hartney, M. T., & Finger, L. K. (2020). Politics, markets, and pandemics: Public education’s
response to COVID-19 (Working Paper No. ai20-304). Brown University.
doi.org/10.26300/8ff8-3945
Holman, M. R., Farris, E. M., & Summer, J. L. (2020). Local political institutions and first-
mover policy responses to COVID-19 (Working Paper). Retrieved from
https://www.dropbox.com/s/8vk25vu6ixsel32/Holman%20et%20al%20COVID.pdf?dl=0
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 38
Hopkins, D. J. (2018). The increasingly United States: How and why American political
behavior nationalized. Chicago, IL: University of Chicago Press.
Hovey, H. A. (1999). State spending for higher education in the next decade: The battle to
sustain current support. San Jose, CA: National Center for Public Policy and Higher
Education.
Hu, L. T., & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to
underparameterized model misspecification. Psychological Methods, 3(4), 424.
doi:10.1037//1082-989x.3.4.424
Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology: A social identity perspective on
polarization. Public Opinion Quarterly, 76(3), 405-431. doi:10.1093/poq/nfs038
Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N., & Westwood, S. J. (2019). The Origins
and Consequences of Affective Polarization in the United States. Annual Review of
Political Science, 22(1), 129-146. https://doi.org/10.1146/annurev-polisci-051117-
073034
Jaquette, O., & Curs, B. R. (2015). Creating the out-of-state university: Do public university
increase nonresident freshman enrollment in response to declining state appropriations.
Research in Higher Education, 56, 535-565. doi:10.1007/s1162-015-9362-2
Jones, J. M. (2020, January 21). Trump third year sets new standard for party polarization.
Washington, DC: Gallup. Retrieved from https://news.gallup.com/poll/283910/trump-
third-year-sets-new-standard-party-polarization.aspx
Kenny, D. A., & McCoach, D. B. (2003). Effect of the number of variables on measures of fit in
structural equation modeling. Structural Equation Modeling, 10(3), 333-351.
doi:10.1207/S15328007SEM1003_1
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 39
Kettl, D. F. (2020). States divided: The implications of American federalism for COVID-19.
Public Administration Review, 80(4), 595-602. doi:10.111/puar.13243
Klem, L. (2000). Structural equation modeling. In L. G. Grimm & P. R. Yarnold (Eds.), Reading
and understanding MORE multivariate statistics (pp. 227-260). Washington, DC:
American Psychological Association.
Kline, R. B. (2012). Assumptions in structural equation modeling. In R. H. Hoyle (Ed.),
Handbook of structural equation modeling (pp. 111–125). New York, NY: Guilford.
Kramer, S. (2020, August 27). More Americans say they are regularly wearing masks in stores
and other businesses. Washington DC: Pew Research Center.
https://www.pewresearch.org/fact-tank/2020/08/27/more-americans-say-they-are-
regularly-wearing-masks-in-stores-and-other-businesses/
Krupenkin, M. (2020). Does partisanship affect compliance with government recommendations?
Political Behavior. doi:10.1007/s111109-020-09613-6
Kwan, J. L., & Chan, W. (2011). Comparing standardized coefficients in structural equation
modeling: A model reparameterization approach. Behavior Research Methods, 43(3),
730-745. https://doi.org/10.3758/s13428-011-0088-6
Lederman, D. (2020, August 6). COVID-19 roundup: Colleges shift strategy in waves. Inside
Higher Education. Retrieved from
https://www.insidehighered.com/news/2020/08/06/covid-19-roundup-fall-strategy-shifts-
waves-zero-mortality-risk-students
Li, C. H. (2016). Confirmatory factor analysis with ordinal data: Comparing robust maximum
likelihood and diagonally weighted least squares. Behavior Research Methods, 48(3),
936-949. doi:10.3758/s13428-015-0619-7
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 40
Looney, A., & Lee, V. (2019). Does the 90/10 rule unfairly target proprietary institutions or
under-resourced schools? Washington, DC: Brookings Institute. Retrieved from
https://www.brookings.edu/research/does-the-90-10-rule-unfairly-target-proprietary-
institutions-or-under-resourced-schools/
Paltiel, A. D., Zheng, A., & Walensky, R. P. (2020). Assessment of SARS-CoV-2 screening
strategies to permit the safe reopening of college campuses in the United States. JAMA
Network Open, 3(7). doi:e2016818-e2016818
Perna, L. W., & Finnery, J. E. (2014). The attainment agenda: State policy leadership in higher
education. Johns Hopkins University Press, Baltimore: MD.
Mangrum, D., & Niekamp, P. (2020). JUE Insight: College student travel contributed to local
COVID-19 spread. Journal of Urban Economics, 103311.
https://doi.org/10.1016/j.jue.2020.103311
Manzoor, S. R., Ho, J. S. Y., & Al Mahmud, A. (2020). Revisiting the ‘university image model’
for higher education institutions’ sustainability. Journal of Marketing for Higher
Education, doi:10.1080/08841241.2020.1781736
Marsicano, C. R., & Brooks, C. (2020). Professor Smith Goes to Washington: Educational
Interest Group Lobbying, 1998–2017. Educational Researcher, 49(6), 448-453.
https://doi.org/10.33774/apsa-2019-hxn00
Marsicano, C., Felten, K., Toledo, L., & Buitendorp, M. (2020). Tracking Campus Responses to
the COVID-19 Pandemic (Working Paper). American Political Science Association,
Washington, DC. doi:10.33774/apsa-2020-3wvrl
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 41
Marsicano, C. Koo, D., Rounds, E., & Owusu, S. (Forthcoming - 2021) “College and University
COVID-19 Testing Protocols by Mode of Instruction— United States, Fall 2020” CDC
MMWR.
Mason, L. (2018). Uncivil agreement: How politics became our identity. Chicago, IL: University
of Chicago Press.
Masserini, L., Bini, M., & Pratesi, M. (2019). Do quality of services and institutional image
impact students’ satisfaction and loyalty in higher education? Social Indicators Research,
146(1), 91-115. doi:10.1007/s11205-018-1927-y
McLendon, M. K., Hearn, J. C., & Mokher, C. G. (2009). Partisans, professionals, and power:
The role of political factors in state higher education funding. The Journal of Higher
Education, 80(6), 686-713. https://doi.org/10.1080/00221546.2009.11779040
Menasce-Horowitz, J. (2020). Republicans, Democrats differ over factors K-12 schools should
consider in deciding whether to reopen. Washington, DC: Pew Research Center.
Retrieved from https://www.pewresearch.org/fact-tank/2020/08/05/republicans-
democrats-differ-over-factors-k-12-schools-should-consider-in-deciding-whether-to-
reopen/
Miller, G. N. S., & Morphew, C. C. (2017). Merchants of optimism: Agenda-setting
organizations and the framing of performance-based funding for higher education. The
Journal of Higher Education, 88(5), 754–784. doi:10.1080/00221546.2017.1313084
Muller, S. (1987). Should states support private colleges – Yes! In L. E. Goodall (Ed.), When
colleges lobby states: The higher education/state government connection (pp. 162-166).
Washington, DC: American Association of State Colleges and Universities.
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 42
National Center for Education Statistics. (2020). IPEDS: Integrated Postsecondary Education
Data System. Washington, DC: Author.
https://nces.ed.gov/ipeds/datacenter/DataFiles.aspx?goToReportId=7
National Conference of State Legislatures (2020). State partisan composition. Washington, DC:
Author. Retrieved from https://www.ncsl.org/research/about-state-legislatures/partisan-
composition.aspx#
National Public Radio. (2020). States are broke and many are eyeing massive cuts. Here’s how
yours is doing. Retrieved from https://www.npr.org/2020/08/03/893190275/states-are-
broke-and-many-are-eyeing-massive-cuts-heres-how-yoursis-doing
New York Times, The. (2020a, December 15). Election night 2020 results. Retrieved from
https://github.com/MEDSL/election_night2020/blob/main/results/nyt_counties_final1204
2020.csv
New York Times, The. (2020b, December 20). Tracking the Coronavirus at U.S. colleges and
universities. Retrieved from https://www.nytimes.com/interactive/2020/us/covid-college-
cases-tracker.html
Nguyen, S., Fishman, R., Weeden, R., Harnisch, T. (2020). The impact of COVID-19 on state
higher education budgets. Washington, DC: New America. Retrieved from
https://www.newamerica.org/education-policy/reports/state-budget-cuts/#authors
Parker, K., Barroso, A., & Fry, R. (2020, October 26). Americans are divided on whether
colleges that brought students back to campus made the right decision. Washington, DC:
Pew Research Center. Retrieved from https://www.pewresearch.org/fact-
tank/2020/10/26/americans-are-divided-on-whether-colleges-that-brought-students-back-
to-campus-made-the-right-decision/
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 43
Parker, K., (2019, August 19). The growing partisan divide in view of higher education.
Washington, DC: Pew Research Center. Retrieved from
https://www.pewsocialtrends.org/essay/the-growing-partisan-divide-in-views-of-higher-
education/
Pink-Harper, A. S. (2018). Does county form of government impact economic growth and
development trends? A case of four states. The American Review of Public
Administration, 48 (3), 245-259. https://doi.org/10.1177/0275074016682314
Oberhauser, A. M., Krier, D., & Kusow, A. M. (2019). Political moderation and polarization in
the heartland: Economics, rurality, and social identity in the 2016 U.S. presidential
election. The Sociological Quarterly, 60(2), 224-244.
doi:10.1080/00380253.2019.1580543
Okunade, A. A. (2004). What factors influence state appropriations for public higher education
in the United States? Journal of Education Finance, 30(2), 123-138.
Richardson, R. C., Bracco, K. R., Callan, P. M., & Finney, J. E. (1999). Designing state higher
education systems for a new century. Phoenix, AZ: ACE and The Oryx Press.
Richmond, C. S., Sabin, A. P., Jobe, D. A., Lovrich, S. D., & Kenny, P. A. (2020). SARS-CoV-2
sequencing reveals rapid transmission from college student clusters resulting in
morbidity and deaths in vulnerable populations (Working Paper). medRxiv.
https://doi.org/10.1101/2020.10.12.20210294
Singh, P. (2021, January 27). ‘I would like my normal lungs back’: Students recount physical
and emotional toll of COVID-19. The Daily Pennsylvanian. Retrieved from
https://www.thedp.com/article/2021/01/penn-students-covid-symptoms-pandemic
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 44
Tandberg, D. A. (2010). Politics, interest groups and state funding of public higher education.
Research in Higher Education, 51(5), 416-450. https://doi.org/10.1007/s11162-010-
9164-5
Tay, K. (2018). County-level data for 2016 presidential elections. Stanford University. Retrieved
from https://web.stanford.edu/~kjytay/courses/stats32-
aut2018/Session%206/Elections_analysis.html
Taylor, B. J., Cantwell, B., Watts, K., & Wood, O. (2020). Partisanship, white racial resentment,
and state support for higher education. The Journal of Higher Education, 91(6), 858-887.
10.1080/00221546.2019.1706016
Tellis, G. J., Sood, N., & Sood, A. (2020). Why did US governors delay lockdowns against
COVID-19? Disease science vs. leaning, cascades, and political polarization (Working
Paper No. 3575004). SSRN. Retrieved from
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3575004
Tollefson, T. A. (2009). Community college governance, funding, and accountability: A century
of issues and trends. Community College Journal of Research and Practice, 33(3-4), 386-
402. doi:10.1080/10668920802580481
Treisman, R. (2020, August 17) UNC-Chapel Hill classes to move online after 130 students test
positive in first week. National Public Radio. Retrieved from
www.npr.org/sections/coronavirus-live-updates/2020/08/16/903071127/less-than-a-
week-after-starting-classes-unc-chapel-hill-reports-4-covid-19-clust.
Tyson, A. (2020, July 22). Republicans remain far less likely than Democrats to view COVID-19
as a major threat to public health. Washington, DC: Pew Research Center. Retrieved
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 45
from https://www.pewresearch.org/fact-tank/2020/07/22/republicans-remain-far-less-
likely-than-democrats-to-view-covid-19-as-a-major-threat-to-public-health/
USAFacts. (2020). U.S. Coronavirus cases and deaths. Retrieved from
https://usafacts.org/visualizations/coronavirus-covid-19-spread-map/
U.S. Census Bureau. (2020). American Community Survey 5-year estimates. Retrieved from
https://factfinder.census.gov/faces/nav/jsf/pages/download_center.xhtml
Whitford, E. (2020, July 7). International students banned from online-only instruction. Inside
Higher Ed. Retrieved from
https://www.insidehighered.com/news/2020/07/07/department-homeland-security-rule-
bans-international-students-online-only
Wofford, A. M. (2020). Modeling the Pathways to Self-Confidence for Graduate School in
Computing. Research in Higher Education, 1-33. https://doi.org/10.1007/s11162-020-
09605-9
Xia, Y., & Yang, Y. (2019). RMSEA, CFI, and TLI in structural equation modeling with ordered
categorical data: The story they tell depends on the estimation methods. Behavior
Research Methods, 51(1), 409-428. doi:10.3758/s13428-018-1055-2
Yong, A. G., & Pearce, S. (2013). A beginner’s guide to factor analysis: Focusing on exploratory
factor analysis. Tutorials in Quantitative Methods for Psychology, 9(2), 79-94.
doi:10.20982/tqmp.09.2.p079
Zingher, J. N. (2018). Polarization, demographic change, and White flight from the Democratic
Party. The Journal of Politics, 80(2), 860-872. doi:10.1086/69699
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 46
Tables
Table 1
Correlation Matrix of Variables in Reported Model
1
2
3
4
5
6
7
1
Primarily In-Person
2
Republican State Control
.10***
3
% of State without Bachelor’s Degree
.06**
.55***
4
’16 County Vote to GOP President Candidate
.16***
.33***
.35***
5
% of County without Bachelor’s Degree
.05*
.24***
.38***
.58***
6
14-Day Average State COVID Cases (Per Capita)
-.04*
.34***
.22***
.07***
.06*
7
14-Day Average County COVID Cases (Per Capita)
-.06**
.27***
.18***
-.02
.08***
.68***
p≤.10+, p≤.05*, p≤.01**, p≤.001***
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 47
Table 2
Model Goodness and Badness of Fit Statistics
Overall Model
4-Year Public
4-Year Private
2-Year Public
CFI
.98
.99
.98
.99
TLI
.96
.98
.96
.99
RMSEA
.05
.03
.06
.05
SRMS
.03
.03
.04
.04
Table 3
Observed and Latent Variable R2
Overall Model
4-Year Public
4-Year Private
2-Year Public
Observed Variables
State Republican Control
.59
.59
.63
.62
State % Without Bachelor’s or Higher
.50
.44
.51
.49
2016 Share of County for GOP Candidate
.77
.59
.73
.75
County % Without Bachelor’s or Higher
.50
.24
.51
.55
State COVID Cases Per Capita (100k)
.85
.83
.94
.72
County COVID Cases Per Capita (100k)
.54
.49
.58
.52
State Revenue Declines
.17
.13
.17
.16
In-Person Instruction
.06
.05
.08
.04
Endogenous Latent Constructs
County Sociopolitical Features
.30
.40
.40
.28
Pandemic Severity
.23
.29
.17
.24
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 48
Table 4
Structural Equation Modeling Examining Influences on Institutions Choosing to be In-Person as Classes Started
Overall
Model
4-Year Public
Institutions
4-Year Private
Institutions
2-Year Public Institutions
Direct
Indirect
Total
Direct
Indirect
Total
Direct
Indirect
Total
Direct
Indirect
Total
County Features
State Features
.56***
.57***
.64***
.64***
.63***
.63***
.53***
.53***
Pandemic Severity
State Features
.53***
-.13***
.40***
.69***
-.23*
.46***
.53***
-.21***
.32***
.54***
-.07***
.47***
County Features
-.23***
-.23***
-.36**
-.36***
-.33***
-.33***
-.13*
-.13*
State Revenue Changes
State Features
.36***
.03**
.38***
.37***
-.01
.36***
.31***
.08+
.39***
.36***
-.04+
.31***
County Features
.09**
.01*
.10***
.01
.01
.02.
.14**
.01
.15**
.11**
.01
.13**
Pandemic Severity
-.05**
-.05**
-.03
-.03
-.03
-.03
-.08+
-.08+
In-Person Instruction
State Features
.07+
.02
.09***
.15
.01
.17**
-.01
.12*
.11***
-.01
.08***
.07***
County Features
.10***
.03***
.13***
.09
.03
.12
.24***
.04**
.28***
.22***
.01
.23***
Revenue Declines
.04+
.04+
.02
.02
.07*
.07**
.00
.00
Pandemic Severity
-.10***
-.00
-.10***
-.07
-.00
-.07
-.09**
-.00
-.09*
-.07
.00
-.07
CFI
.98
.99
.98
.99
TLI
.96
.98
.96
.99
RMSEA
.05
.03
.06
.05
SRMS
.03
.03
.04
.04
p≤.10+, p≤.05*, p≤.01**, p≤.001***
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 49
Figure 1 - Hypothesized Model to Test
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 50
Figure 2 - Accepted Model. Direct Robust Standardized Coefficients Reported, Solid Lines Denote Significant Relationships, Dashed Lines Are Non-Significant
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 51
Appendix
Table A1
Models with Only Observed Share of GOP Vote in 2016 at the County Level
Overall
Model
4-Year Public
Institutions
4-Year Private
Institutions
2-Year Public Institutions
Direct
Indirect
Total
Direct
Indirect
Total
Direct
Indirect
Total
Direct
Indirect
Total
County Features
State Features
.45***
.45***
.47***
.47***
.50***
.50***
.42***
.42***
Pandemic Severity
State Features
.49***
-.09***
.41***
.57***
-.11***
.46***
.47***
-.14***
.33***
.51***
-.04*
.47***
County Features
-.18***
-.18***
-.24***
-.24***
-.28***
-.28***
-.10*
-.10*
State Revenue Changes
State Features
.36***
.02
.38***
.35***
.02
.36***
.34***
.05*
.39***
.36***
-.04+
.32***
County Features
.11***
.01*
.11***
.05
.00
.06
.14***
.01
.15***
.13***
.01
.14***
Pandemic Severity
-.05*
-.05*
-.02
-.02
-.03
-.03
-.08+
-.08+
In-Person Instruction
State Features
.08*
.02
.10***
.18*
.00
.18***
.04
.08*
.12**
.02
.06***
.08***
GOP Voteshare
.12***
.02***
.15***
.09+
.02
.11*
.21***
.04**
.25***
.19***
.01
.20***
Revenue Declines
.03
.03
.01
.01
.07*
.07*
.00
.00
Pandemic Severity
-.10***
-.01
-.10***
-.08
.00
-.08+
-.10**
-.00
-10**
-.07
.00
-.07
CFI
.98
.99
.99
.99
TLI
.96
.99
.96
.99
RMSEA
.05
.01
.06
.04
SRMS
.03
.02
.03
.03
p≤.10+, p≤.05*, p≤.01**, p≤.001***
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 52
Figure A1 - Technical Layout of SEM Underlying Structure
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
WE WANT YOU BACK 53
Figure A2 - Alternative Model with Prestige and Residential Characteristics – Unacceptable Fit Statistics
This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=3778772
