Download as pdf or txt
Download as pdf or txt
You are on page 1of 28

POLICY BRIEF

COVID-19 and Michigan’s


Lowest Performing Schools
Emily Hatch and Erica Harbatkin

May 2021

EPIC
Education Policy
Innovation Collaborative
RESEARCH WITH CONSEQUENCE
ACKNOWLEDGEMENTS
The authors wish to acknowledge the many people who provided technical and research support for this brief including Melissa
Lovitz, Amy Cummings, Jesse Nagel, Nicole Rausch, Lauren Schoen, Mallory Weiner, Gabrielle Wihongi, Jeremy Anderson and
Jesse Nagel. We are also indebted to our colleagues at the Michigan Department of Education and the Center for Educational
Performance and Information for their valuable partnership in this research, in particular Roderick Bernosky, Josh DeBradbander,
Tom Howell, Carl Jones, Joshua Long, Michael McGoarty, Holly Willson, William Pearson, and Gloria Chapman. Finally, we thank
Bridgette Redman for her work copy editing, and Adam K. Edgerton of Learning Policy Institute for his insightful feedback on an
earlier version of this brief.

DISCLAIMER
The Education Policy Innovation Collaborative (EPIC) at Michigan State University is an independent, non-partisan research center
that operates as the strategic research partner to the Michigan Department of Education (MDE) and the Center for Educational
Performance and Information (CEPI). EPIC conducts original research using a variety of methods that include advanced statistical
modeling, representative surveys, interviews, and case study approaches. This research used data structured and maintained
by the MERI-Michigan Education Data Center (MEDC). MEDC data are modified for analysis purposes using rules governed by
MEDC and are not identical to those data collected and maintained by MDE and/or CEPI. Results solely represent the analysis,
information, and opinions of the author(s) and are not endorsed by and do not reflect the views or positions of grantors, MDE and
CEPI, or any employee thereof.
EPIC
Education Policy
Innovation Collaborative
RESEARCH WITH CONSEQUENCE

May 2021

COVID-19 and Michigan’s


Lowest Performing Schools
By Emily Hatch and Erica Harbatkin

INTRODUCTION
As the world has grappled with the COVID-19 pandemic, educators and families have made
extraordinary efforts toward supporting and teaching children. Educators have developed and
implemented creative solutions to providing remote education and districts have drawn on
available resources to hire mental health counselors, provide students with needed technology,
and support teachers implementing remote learning. They have made these efforts against the
backdrop of a health crisis that has killed more than half a million Americans and more than 3
million people worldwide, undermined economic conditions, and isolated people from family
and friends for more than a year.

Students in the nation’s lowest performing schools have been doubly affected by the COVID-19
pandemic. Early evidence suggests that students in communities and schools that were already
underserved and identified as low performing before the pandemic are both more likely to
be learning remotely1 and to be struggling academically during the pandemic.2 Moreover, the
COVID-19 pandemic has taken an especially heavy toll on communities with high rates of poverty
and underrepresented minorities—the same communities that are home to many of the low-
performing schools that states have targeted for school and district turnaround. As job losses
have mounted in these communities, families have reduced spending on food; estimates suggest
that 17 million new Americans will become food insecure as a result of the pandemic.3 Lower
EPIC | Education Policy Innovation Collaborative — Michigan State University

income households with food insecurity face increased risk of chronic disease development and
complications and higher mortality.4 Thus, these low-performing schools nave a particularly
heavy lift in their efforts to educate students during the pandemic and ultimately to accelerate
student learning once schools return to their new normal after the pandemic wanes.

This policy brief explores the pandemic’s effects on the districts and students served by
Michigan’s Partnership Model of School and District Turnaround. Michigan first implemented
the Partnership Model in 38 chronically low-performing
schools in nine districts in the 2016-17 school year, and
Students in then identified two additional rounds for a total of 86 new

communities and turnaround schools in 27 districts in the 2017-18 school


year. Since the beginning of Partnership, nine districts have

schools that were either exited or closed, leaving a total of 99 Partnership


schools housed within 27 Partnership districts in Michigan.
already underserved These 27 districts represent just over 3% of the state’s
districts. Partnership districts are themselves housed
and identified as low- within 11 counties in the state (13%). In this policy brief, we
provide background and context on the communities where
performing before Partnership districts are located, highlight the strategies

the pandemic are Partnership districts are using to educate their students
during the pandemic and how those strategies differ from

more likely to be other districts, and discuss the implications for policy and
practice related to low-performing schools.
learning remotely We show that Partnership communities are home to a

and to be struggling disproportionate share of the state’s underrepresented


minorities, have higher rates of poverty and food insecurity
academically during than other communities throughout the state, experienced
extraordinarily high rates of COVID-19 in the early days of
the pandemic. the pandemic, and have suffered more COVID-19 deaths
than non-Partnership communities. Educators responded
to these challenges by providing students with food, on-site mental health counselors, and
technology resources—though there is still work to be done to close the opportunity gap for all
students in these districts. We conclude this policy brief with some considerations for supporting
Partnership districts and schools in light of these findings as the threat of COVID-19 wanes and
educators work to accelerate student learning in future school years.

SAMPLE AND DATA


Table 1 shows the data sources, measures of interest we drew from each source, date from
which the data were drawn, and sample size of Partnership districts and counties from which
we have each measure.

4
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

TABLE 1. Data Sources


Data Measures of interest Source Date(s) Sample size

Community Race, ethnicity, U.S. Census 2018 83 counties


characteristics poverty, educational Bureau American (11 with
attainment, Community Partnership
internet access, Survey (ACS) five- districts)
computer and device year estimates
access, receipt
of Supplemental
Nutrition Assistance
Program (SNAP)
benefits
District revenue Federal, state, local, Michigan 2019-20 820 Michigan
data and other revenues Department of school districts
Education (MDE) with reported
and the Center revenue
for Educational (27 Partnership)
Performance
and Information
(CEPI)
District plans District plans Continuity of Spring 2020 813 COL plans
to continue for provision of Learning (COL)
student technology, meals, plans
learning in and mental health
spring 2020 services and
during the resources during
COVID-19 COVID-19
pandemic
Monthly District monthly Reconfirmed Monthly, 799-814 districts
instructional plans to educate Extended September per month
modality plans students fully in- COVID-19 2020 through (26 Partnership)*
person, fully remote, Learning (ECOL) March 2021
or in hybrid format Plan Monthly
Questionnaire
COVID-19 case, County-level Michigan Daily, 83 counties
death, and COVID-19 cases and Department April 1, 2020 (11 with
positivity rates deaths per 100,000 of Health and through Partnership
population, test Human Services March 3, 2021 districts)
positivity rates (MDHHS)
* One Partnership district, Insight School of Michigan, a fully virtual district, is not included in analyses from these
plans because it was remote prior to the pandemic.

5
EPIC | Education Policy Innovation Collaborative — Michigan State University

As we show in Table 1, we draw from county-level data collected by the U.S. Census, data
on school district revenue sources, district-submitted plans for serving students during the
pandemic, monthly instructional modality plans, and county-level data on COVID-19 cases,
deaths, and positivity rates over time. We provide additional detail on each of these data
sources and how we constructed measures in the data appendix.

METHODS
We analyzed all data at the district level. We assigned county-level values from the U.S.
Census Bureau American Community Survey (ACS) da and COVID-19 rates to districts, and
then calculated means for Partnership and non-Partnership districts. In most analyses, we
weighted these means by district size (i.e., number of students in the district). As a result, the
means can be interpreted as the experience of the average student in a Partnership or non-
Partnership district. In cases where the weighted means are substantively different from the
simple unweighted means, we also show the simple means. These simple means provide the
average district experience.

Because urbanicity is associated with access to technology, we ran separate analyses by Census
locale code for the technology-related ACS variables, collapsing all districts coded as city (large,
midsize, or small) into a single urban category, and all districts coded as suburb or town into a
single suburb/town category. Because only one Partnership district is rural, we focused these
analyses on comparisons within these urban and suburb/town categories. Examining technology
variables by Partnership status within locale allows us to focus on meaningful differences across
districts with similar urbanicity.

FINDINGS
Finding 1: Counties with Partnership Districts Are Home to a
Disproportionate Share of Underrepresented Minorities and Have
Higher Rates of Poverty and Lower Median Incomes Than Other
Counties in the State
Figure 1 provides the race (top bars) and ethnicity (lower bars) breakdown of counties where
Partnership and non-Partnership districts are located. The size of the dark blue bars in particular
show that Partnership communities are home to a disproportionate percentage of Black
residents, with more than twice as many residents in Partnership communities identifying as
Black than in non-Partnership communities (27% vs 13%). The second set of bars shows that
Partnership students also live in counties with higher rates of Latinx residents than do non-
Partnership students.

6
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

FIGURE 1. County Race and Ethnicity by Presence of Partnership Districts


2.9% 5.1%

Partnership 64.5% 27.4%

3.1% 4.6%

Non-Partnership
79.3% 13.0%

0 10 20 30 40 50 60 70 80 90 100
Percent of Residents

6.2%
Partnership
White Black
Asian Other
Hispanic or Latinx (any race)
Non-Partnership
5.0%

0 10 20 30 40 50
Percent of Residents

Note: Data from 2018 ACS five-year county estimates. Figures represent student-weighted means. The Hispanic or
Latinx ethnicity category is separate from race (e.g., a person who is White and Hispanic or Latinx would appear in
both the race bars and the Hispanic or Latinx bars) due to ACS reporting.

The effects of racial segregation are especially salient as the COVID-19 pandemic has
highlighted health inequities throughout the country, with Black, Latinx, and Indigenous
populations dying at twice the rate of the White population.5 Additionally, Black and Latinx
renters and homeowners are facing more challenges making rent and mortgage payments
during the pandemic, and a larger share of Black and Latinx adults report that they are worse
off financially since the pandemic began.6 It is clear from these demographics that the damage
done by the pandemic will fall disproportionately on families living in Partnership communities.

Figure 2 shows that the median household income is 12.5% lower and the individual poverty
rate is about one-third higher in Partnership than non-Partnership communities. Poverty is
especially deep-seated in households with children; the family and child poverty rates are
about 40% higher in Partnership than in non-Partnership counties. In total, 28% of children
in Partnership communities live in poverty. During the pandemic, the effects of poverty have
become even more stark, with families who were lower income at the start of the pandemic
more likely to suffer job losses and to report being worse off financially than families with
higher incomes at the start of the pandemic.7 These grim statistics underscore the increasing

7
EPIC | Education Policy Innovation Collaborative — Michigan State University

challenges Partnership districts face in supporting children who may be struggling with
housing and food insecurity.

FIGURE 2. County Socioeconomic Characteristics by


Presence of Partnership Districts

Community Characteristics
Partnership •vs• Non-Partnership Districts
30 30
28.2%
25 25
Percent of Residents

Percent of Residents
20 20
19.6% 20.2%
15 15
14.5% •vs• 14.7%
10 10
10.3%
5 5

0 0

Partnership Non-Partnership

Family Poverty Rate Individual Poverty Rate Under 18 Poverty Rate

Median
$51,728 Household
Income
$59,146
Note: Data from 2018 ACS 5-year county estimates. Figures represent student-weighted means.

Finding 2: COVID-19 Hit Partnership Communities Earlier


and Harder Than Non-Partnership Communities
Figure 3 shows seven-day rolling average COVID-19 case positivity rates (i.e., the percentage
of COVID-19 tests that came back positive) over time for Partnership (green solid line) and
non-Partnership (dark blue dotted line) communities, highlighting two trends. First, positivity
rates were about 60% higher in Partnership communities than non-Partnership communities
in the early days of the pandemic—with more than 45% of early tests yielding positive
results compared with less than 30% in non-Partnership communities. Second, positivity

8
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

rates remained higher in Partnership than non-Partnership communities through the 2019-20
school year before the rates converged in early fall 2020.

FIGURE 3. Case Positivity Rate Over Time by Partnership Status

40
40

30
30
Positivity rat e
Positivity Rate

20
20

10
10

0
0

Apr20 May20 Jun20


Apr May JunJul20 JulAug20AugSep20SepOct20OctNov20NovDec20DecJan21JanFeb21Feb
Mar21 Apr21 May21
Mar Apr
2020
Non-Partnershi p Partnershi p 2021
Partnership Non-Partnership

NOTE: County-level test positivity rates up to May 3, 2021. In an earlier version of this report, we erroneously
concluded that positivity rate remained higher in Partnership districts by the beginning of fall 2021. This is not the
case, and has been corrected here.
N=826 districts

The early surge in Partnership communities is also


evident in Figure 4, which shows seven-day rolling
Positivity rates
average case (first panel) and death (second panel) rates
per 100,000 population over time in Partnership and
were nearly 60%
non-Partnership communities. Partnership communities higher in Partnership
experienced more than 20 cases per 100,000 residents
per day by early April 2020 while non-Partnership communities than
communities experienced about half as many cases. By
mid-April 2020, death rates in Partnership communities in non-Partnership
had risen to 2.5 per 100,000 population—again twice
the rate of non-Partnership communities. Partnership
communities in the
and non-Partnership communities equalized over time
as new public safety measures took effect, resulting in
early days of the
similar case and death rates by summer and through pandemic.
the 2020-21 school year.

9
EPIC | Education Policy Innovation Collaborative — Michigan State University

FIGURE 4. Case and Death Rates Over Time by Partnership Status

80
80

70

60
60

50
per100K
Casesper 100K
40

40
Cases

30

20
20

10

0
0

Mar15 Apr15 May15 Jun15 Jul15 Aug15 Sep15 Oct15 Nov15 Dec15 Jan15 Feb15 Mar15 Apr15May1
Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr
2020
Non-Partnershi p Partnershi p 2021
Partnership Non-Partnership

3
3

2
2
per100K
Deathsper
Deaths 100K

1
1

0
0

Mar15 Apr15 May15 Jun15 Jul15 Aug15 Sep15 Oct15 Nov15 Dec15 Jan15 Feb15 Mar15 Apr15May1
Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr
2020
Non-Partnershi p Partnershi p 2021
Partnership Non-Partnership
Note: County-level cases and deaths beginning March 6, 2020 and ending April 26, 2021, applied to districts.
N=826 districts

In the aggregate, although infections have evened out across Partnership and non-Partnership
communities, Figure 5 shows that the difference in cumulative death rate is stark. While cumulative
case rates are only about 4% higher in Partnership counties, cumulative death rates are 29%

10
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

higher. That means students in Partnership districts are substantially more likely to have lost a
friend or family member to COVID-19 at some point during the pandemic than students in non-
Partnership districts. Dividing the death rate by the case rate points to additional inequalities—a
higher death-to-case rate may stem from a number of factors, including health disparities (e.g.,
health insurance, access to high quality health care), economic conditions (e.g., ability to take
time off work), and others. Specifically, the death-to-case rate is 2.6% in Partnership and 2.1% in
non-Partnership communities.

FIGURE 5. Cumulative Case and Death Rates per 100,000 by Partnership Status

Partnership 8,290 218

Non-Partnership 7,970 169

0 1K 2K 3K 4K 5K 6K 7K 8K 9K 0 50 100 150 200 250


Cumulative Cases Per 100K Cumulative Deaths Per 100K

Partnership Communities Non-Partnership Communities

Note: County cases and deaths as of April 26, 2021, applied to districts. Weighted by district size.
N=826 districts

Finding 3: While All Districts Planned to Address Student Mental


Health Needs, Non-Partnership Districts Were More Likely to Provide
On-Site Access to Mental Health Counselors
Students experiencing more exposure to illness and death may also grapple outside of school
with the mental health implications and trauma of COVID-19. Figure 6 shows district plans
for providing mental health resources by Partnership status. Here, we show both unweighted
(district level, left side of graph) and weighted (student level, right side of graph) figures
separately to underscore important differences. While the first set of bars shows that almost
all districts cited plans to address mental health needs in their Continuity of Learning (COL)
plans, the ways in which they planned to deliver these services—along with the scope—varied
by Partnership status. The left side of the graph shows that more Partnership districts (50%)
planned to provide mental health resources to all students than non-Partnership districts
(38.5%). But the right side of the graph shows that these patterns are reversed at the student
level—with only 35% of students in Partnership districts compared with 47% of students in
non-Partnership districts that provide mental health services to all students. In other words,
the Partnership districts that planned to provide universal mental health services are smaller
than the Partnership districts that planned to provide these services to only some students,

11
EPIC | Education Policy Innovation Collaborative — Michigan State University

whereas rates were similar for smaller and larger non-Partnership districts. This means
that, although more Partnership districts provided universal resources than non-Partnership
districts, Partnership students were less likely than their non-Partnership peers to be in a
district that did so. This finding points to challenges of scale that may be especially salient for
Partnership districts—while smaller Partnership districts were able to provide socioemotional
or mental health resources to all students, larger districts planned to provide these resources
to a subset of students.

The next set of bars in the district-level panel on the left side shows that more than 8 in 10
non-Partnership districts planned to provide access to on-site counselors or mental health
professionals, compared with just 6 in 10 Partnership districts. The student-level panel on
the right side shows an even wider gap, with just 4.5 in 10 Partnership students in districts
benefiting from district-provided on-site counselors—about half the frequency of their non-
Partnership peers. Put another way, while 60% of Partnership districts planned to provide
on-site counselors, only 45% of students in Partnership

While 60% of districts have access to on-site mental health counselors.


That divide does not exist in non-Partnership districts,
Partnership districts where just over 80% of districts planned to provide on-
site counselors and just over 80% of students had access
planned to provide to counselors. Again, these differences suggest that the
challenges with providing mental health resources to
on-site counselors, larger populations of students may be especially stark in
Partnership districts.
only 45% of students
Instead, Partnership districts planned to leverage the
in Partnership resources they already had. The third set of bars shows
district plans to ask staff to check in with students
districts have access about socioemotional and mental health. Sixty percent

to on-site mental of Partnership districts planned to ask staff to conduct


check-ins with students about their mental health
health counselors. and 69% of students were in districts that did so. By
comparison, 39% of non-Partnership districts planned to
ask staff to conduct these check-ins and 41% of non-Partnership students were in districts
that did so.

The final set of bars shows a similar pattern, with fewer Partnership districts planning to make
referrals to outside sources (25% in Partnership to 36% in non-Partnership, left panel) but
more Partnership students in districts planning to make these referrals (46% in Partnership
and 38% in non-Partnership, right panel). Together, these differences underscore the resource
constraints in Partnership districts, which may lack sufficient funding for full-time counselors
dedicated to mental health.

12
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

FIGURE 6. Plans to Provide Mental Health Supports by Partnership Status

Socioemotional 50.0% 50.0% 35.1% 64.9%


or mental health
resources
60.9% 38.5% 47.4% 52.6%

Access to site 60.0% 44.9%


counselors/mental
health 81.8% 83.3%

Socioemotional/ 60.0% 68.9%


mental health check
39.1% 40.9%

25.0% 46.1%
Referrals to outside
resources 36.1% 38.3%

100 80 60 40 20 0 0 20 40 60 80 100
District Level Student Level
Partnership Non-Partnership
To all students — Partnership To all students — non-Partnership
To some students — Partnership To some students — non-Partnership

Note: Data from district Continuity of Learning plans submitted in April 2020.
N=813 districts with available COL plans.

Finding 4: Households in Partnership Communities Were More


Likely to Have Food Insecurity Than Households in Non-Partnership
Communities, but Partnership Districts Were Less Likely to
Distribute Meals Than Were Non-Partnership Districts
Food insecurity has been a growing crisis during the pandemic, with likely long-lasting
implications for children who are affected by inadequate access to food. Figure 7 shows that
even before the pandemic began, nearly one-third of Partnership community households
with children relied on federal assistance to pay for groceries compared with 21% in non-
Partnership communities. For families facing food insecurity, school closures may create extra
barriers to accessing food for students through their schools. While the need for Supplemental
Nutrition Assistance Program (SNAP) benefits is higher in Partnership counties, Partnership
districts were slightly less likely than non-Partnership districts to report in their COL plans that
they planned to distribute meals. Specifically, 83% of Partnership students were in districts
that planned to engage in meal distribution, while 97% of non-Partnership students were in
districts with meal distribution plans.

13
EPIC | Education Policy Innovation Collaborative — Michigan State University

FIGURE 7. Federal and District Supports to Address Food Insecurity


100
97%

80
83%

60
Percent

40

30.4%
20
21.6%

0 9.2% 6.3%

Households that receive SNAP Households with children Districts with plans for
that receive SNAP meal distribution

Partnership Non-Partnership

Note: Figures for first two sets of bars calculated from 2018 ACS five-year estimates. County-level figures applied to
districts and weighted by student enrollment. Figures from the last two bars come from district COL plans.

Finding 5: Partnership Districts Have Relied Heavily on Remote-Only


Instruction to Educate Students Throughout the Pandemic
High community spread, death rates, and other community-based factors have led Partnership
districts to rely largely on remote instruction for a large part of the school year. Indeed, Figure
8 shows that more than half of Partnership districts began

In January 2021, the school year fully remote, compared with about 20
percent of non-Partnership districts. Although districts
nearly all Partnership across the state—both Partnership and non-Partnership—
shifted to remote-only instruction in December,
districts were fully Partnership districts continued to plan for remote-only
instruction for January 2021 while the number of non-
remote, compared Partnership districts in remote-only mode decreased

with about 1/3 of non- for the new year. In January 2021, nearly all Partnership
districts were fully remote, compared with about one-

Partnership districts. third of non-Partnership districts. Partnership districts


began to return to in-person and hybrid instruction later
in the 2020-21 school year than other districts in the
state, though some remained remote late into the spring semester.

14
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

FIGURE 8. Instructional Modality Plans by Month


Fully In-Person Option Hybrid
100 100
Percent of Districts Offering Modality

Percent of Districts Offering Modality


80 80

60 60

40 40

20 20

0 0

Sep Oct Nov Dec Jan Feb Mar Apr Sep Oct Nov Dec Jan Feb Mar Apr

Remote Only
100
Percent of Districts Offering Modality

80

60 Partnership
Non-Partnership
40

20

Sep Oct Nov Dec Jan Feb Mar Apr

Note: Markers represent district averages. Fully in-person option means districts have an option for students to
attend in-person for all days. Hybrid classifies districts without a fully in-person option that have any students
attending a hybrid model. Fully remote identifies districts in which all students attend remotely. Figures exclude
virtual districts that were remote prior to the pandemic.

These differences may in fact reflect the preferences of Partnership families, as two national polls
suggest most children were receiving the parents’ preferred mode of instruction.8 Preferences in
Partnership districts may be shaped in part by the underlying disparities of their communities.
Concerns about contracting COVID-19 may be greater in families with higher risk factors, such as
preexisting conditions, inadequate access to care, and lack of paid time off.

Figure 9 sheds some light on those preferences. While Figure 8 shows the number of districts
offering each modality, Figure 9 shows the estimated number of students who actually enrolled
in each instructional modality. Districts were asked to specify the approximate percentage of
students receiving each modality each month (i.e., less than 24%, 25-49%, 50-74%, 75-99%,
and 100%). We combined these responses with district-level student enrollment to estimate the

15
EPIC | Education Policy Innovation Collaborative — Michigan State University

share of all Partnership and non-Partnership students that received each instructional modality.9
The ranges represent the total student shares based on the low (labeled “min uptake”) and high
(labeled “max uptake”) ends of the percentage ranges indicated by each district. The height of the
Offered
light gray bars represents the total percentage of students Max
whose district offered that modality.
Uptake
Offered
Max Uptake Min Uptake
FIGURE 9. Uptake of Instructional Modality by MonthOffered
Only Option
Min Uptake
Max Uptake Partnership
Fully In-Person Only Option
Non-Partnership
100
100 Min Uptake
Partnership Non-Partnership
100 Only Option
8080 Partnership Non-Partnership
In-Person

100
80
6060
Percent
Fully In-Person

80
60
4040
Fully
In-Person

60
2040
Percent

20
Percent

40
Fully

0200
Percent

20
1000
Hybrid
100
1000
80
d

100
80
80
60
d Hybri
d

6080
60
Percent

40
Hybri
Percent

4060
40
Hybri
Percent

20

2040
20
Percent

0200
100

0
100
80
Fully Remote
100
Fully Remote

100
80 60
Fully Remote

8080
60
40
Remote

6060
Percent

40
Percent

20
Percent
Fully

4040
20
0
Percent

September October November December January Februar y March April


20200
September October November December January Februar y March April
00
September October November December January Februar y March April
Sept Oct Nov Dec Jan Feb Mar Apr

Offered

Max Uptake

Min Uptake
Only Option
Partnership Non-Partnership
100
Note: We calculated the percent of students by multiplying district-reported ranges of plans for students to be
80 served by each instructional modality by their total student counts and then summing across all the districts
offering each modality. Figures exclude virtual districts that were remote prior to the pandemic.
Percent Fully In-Person

60

40

16
20

0
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

The difference between the green and blue bars relative to the gray bars shows that across all
months, families in Partnership districts were more likely to take up remote instruction and
less likely to take up in-person instruction even to the extent that it was offered. However, the
upward trend in the first panel combined with the downward trend in the third shows a more
recent shift away from remote and into in-person instruction.

These patterns suggest that both the inequitable provision of in-person instructional options
for students in Partnership relative to non-Partnership districts and the differential take-up
rate of in-person schooling throughout much of the 2020-21 school year could have severe
implications for Partnership district students’ opportunities to learn. Delivering high quality
remote instruction requires access to reliable internet and other needed resources that are less
widely available in Partnership communities. To that end, educators in Partnership districts
face a heavier lift to provide high quality instruction to their students. Moreover, emerging
research from other states suggests that students learn less when engaged in remote relative
to in-person instruction.10 As a result, it is likely that students in Partnership districts will
enter the 2021-22 school year having had less opportunity to learn than their peers in non-
Partnership districts, and will need more intensive supports to accelerate their learning in the
years to come.

Finding 6: Partnership Districts Were More Likely Than Non-


Partnership Districts to Provide Some Forms of Technology Supports
to Make up for Uneven Community Access to Technology—Though
These Supports Were Less Comprehensive in Partnership Districts
The pandemic has highlighted and accentuated the importance of technology in educational
equity as over the past year, many schools have educated their students remotely and
through hybrid instruction. While learning management systems may include mobile settings,
completing assignments on a smartphone or tablet is less practicable than on a laptop.11
Similarly, unreliable or slow internet would inhibit students’ ability to engage with synchronous
instruction, watch videos, or download content. Lack of personal access to a laptop or desktop
computer or inadequate internet access may therefore reduce a student’s opportunity to learn.

These challenges are more pronounced in Partnership communities. ACS data show that
residents of Partnership communities are less likely than residents of non-Partnership
communities to own a desktop or laptop computer, more likely to rely exclusively on a
smartphone or tablet, and less likely to have broadband internet at home. These differences
are larger in non-urban locations.

To address these gaps, districts have needed to find ways to ensure that their students can
access instruction through remote means. They have done so by providing computing devices
and internet access. Figure 10 displays district plans for providing technology. Nearly all
districts—Partnership and non-Partnership—planned to distribute devices to either some or all
of their students, shown by the green bars. However, whereas nearly one-fifth of non-Partnership

17
EPIC | Education Policy Innovation Collaborative — Michigan State University

districts planned to distribute devices to all students (dark green), no Partnership districts
specifically noted their plans to do the same. This finding suggests that while Partnership
districts have a higher level of need to bridge the technological gap for their students, they may
not have the resources to provide devices to all students.

FIGURE 10. Plans for Technology Provision by Partnership Status

94.4% 96.5%

Partnership 45% 28.4%

30% 33.1%

17.8% 77% 78.1% 18.8%

Non-Partnership 28.8% 36.6%

20% 21.8%

100 80 60 40 20 0 0 20 40 60 80 100
District Level Student Level
Devices to Some Students Internet Access or Wi-Fi Hotspot
Devices to All Students Information on Free Internet Access

Note: Data from district Continuity of Learning plans submitted in April 2020.
N=813 districts.

The second two bars in both panels show district plans for providing internet access or
information on free internet access to students. The dark and light blue bars, respectively,
in the left panel show that more Partnership districts (45%) than non-Partnership districts
(29%) planned to provide access to broadband internet and/or Wi-Fi hotspots and to provide
information on how to access free internet access. However, the student-weighted bars in
the right panel show that the average Partnership student

Partnership students was less likely to be in a district providing internet access


(28% Partnership and 37% non-Partnership). Instead,
were less likely to be Partnership students were more likely to be in districts
that provided information on obtaining internet access
in a district providing (33% Partnership and 22% non-Partnership). These
differences may underscore the resource and capacity
internet access. challenges Partnership districts face in providing students
with needed internet access. Providing Wi-Fi hotspots for
a large number of students would require a considerable investment by districts with scarce
resources and high need.

18
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

The differences between Partnership and non-Partnership district technology provision


provide only one dimension of inequity related to technology (i.e., that Partnership students
were less likely to be in a district that provided one-to-one devices or internet access). The
other dimension of inequity is the need for devices and internet, which is greater in Partnership
districts due to two factors. First, device ownership and internet access were less pervasive
in Partnership than non-Partnership communities, as we note above. Second, also discussed
above, Partnership districts spent a larger portion of the school year providing remote-only
instruction, during which personal technology access becomes even more essential.

Finding 7: Partnership Districts Rely Heavily on Revenue From


State and Federal Funding Sources, Which May Be Less Stable
in the Aftermath of the Pandemic
Figure 12 shows total revenue per pupil by Partnership status. Because state and federal funding
is higher for districts with high proportions of low-income students, Partnership districts have
about 16% higher total revenue than non-Partnership districts—though research in Michigan
suggests districts with high concentrations of poverty
would need 39% to 44% more funding per student to
provide their students with an adequate education.12
Research suggests
Additionally, a smaller share of Partnership district districts with high
revenue comes from local property taxes, which tends to
be more stable than other revenue sources. The average concentrations of
Partnership district raises $2,196 per pupil in local
revenue—40% less than the average non-Partnership
poverty would need
district, which raise $3,663 in local revenue. Instead,
39% to 44% more
funding in Partnership districts comes largely from
federal and state sources. Because poverty rates are funding per student
high in Partnership districts, they are eligible for Title I
funding from the federal government, leading to about to provide their
50% more in federal funds—though as Figure 12 shows,
federal dollars make up a very low percentage of total
students with an
revenue. State sources account for the largest share adequate education.
of revenue in both non-Partnership and Partnership
districts—and state dollars are particularly significant in Partnership districts, where two-
thirds of revenue comes from the state. A small portion of these state dollars comes from
21h funds targeted for school turnaround ($6 million to $7 million a year allocated across
all Partnership districts). Local dollars make up only 10% of revenue in Partnership districts
compared with 21% in non-Partnership districts.

19
EPIC | Education Policy Innovation Collaborative — Michigan State University

FIGURE 12. Revenue Sources by Partnership Status


20k Total: $17,686.61
Funding Breakdown
18k
$2,917.97 Total: $15,294.95
8% 10%
16k
14%
$2,483.15
14k $2,249.04

12k $1,137.93
Dollars Per Student

$10,323.75
$8,010.44 68%
10k
Partnership Districts
8k Local Share Federal Share
State Share Other Share
6k

11%
4k 21%
8%
$3,663.43
2k
$2,195.86
0

Partnership Non-Partnership 60%


Local Revenue Per Student Non-Partnership Districts
State Revenue Per Student Local Share Federal Share
Federal Revenue Per Student State Share Other Share
Other Revenue Per Student

Note: Data from Michigan’s Center for Educational Performance and Information (CEPI) public school finance data.
N=820 districts.

In sum, the differences in funding sources underscore important variation in the makeup of
overall funding for Partnership and non-Partnership districts. Partnership districts are unable to
raise the same amount from local funds as non-Partnership districts, and the disparity in funding
is in part—but not completely—equalized through federal monies made available as a result
of the high concentration of poverty in Partnership communities. The remainder of the gap is
filled in through state funds. Both federal and state funding sources are sensitive to politics and
economic downturns, making resource stability a particular challenge for Partnership districts.13
Moreover, these formula-based funding sources are sensitive to student enrollment changes
because they are allocated on a per-pupil basis. To the extent that student enrollment does not
rebound in later school years, COVID-19-related declines will lead Partnership districts to lose
a larger share of their revenue in the absence of continued hold-harmless enrollment provisions
in future school years.

20
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

POLICY IMPLICATIONS
The findings in this policy brief show that the pandemic has exacerbated existing inequalities
in K-12 education in Michigan. The pandemic’s negative effects will continue to reverberate
through communities—and Partnership communities in particular—even after it ends.
Education policy moving forward will need to account for the considerable challenges of
low-performing schools in order to adequately address existing and emerging inequities
underlying school performance. In this final section, we describe policy implications related
to three broad areas—school funding, accountability, and narrowing opportunity gaps.

Even as Schools Return to Traditional In-Person


Learning, There is a Need for Continued and
New Funding to Support School Improvement
Because local funds for schools come from property taxes, variation in community wealth—
in particular, the taxable value of local properties—leads to inequitable funding from local
sources.14 As we show in the previous section, Partnership districts, which have lower property
wealth than non-Partnership districts, therefore need to rely more on state and federal
funds to educate students. Unlike local funds, these state and federal funds are explicitly
tied to enrollment, which has decreased across the country as families seek out alternatives
to public school or children take on non-school responsibilities such as caring for younger
siblings during the pandemic.15 Negative effects of declining enrollment will therefore fall
disproportionately on less affluent schools—like those in Partnership districts—that rely more
on state funds. By contrast, more affluent, higher performing districts get a larger share of
revenue from local property taxes, which will not be affected by enrollment changes. The state
legislature passed measures to mitigate funding shortfalls stemming from lost enrollment for
the 2020-21 school year (specifically, 2020-21 enrollment will be calculated as 75% of 2019-
20 enrollment and just 25% of 2020-21 enrollment16). However, it is unclear whether and how
quickly lost enrollment will rebound, raising concerns about insufficient and uncertain funding
in 2021-22 and beyond.

Previous research on the Partnership Model and on school expenditures more broadly has
found that funding is critical for improvement efforts. Even before the pandemic, Partnership
leaders expressed a need for additional funding to accomplish their improvement goals.17 As
the threat of COVID-19 wanes, businesses continue to reopen, and the economy continues to
recover, students and staff in Partnership districts will be working to return to a new normal.
Nationwide evidence from benchmark assessments during COVID-19 suggests that students
in Partnership districts will need support to accelerate learning as their communities continue
to experience the reverberations from interruptions to learning as well as the loss of family
and friends, lost income, homelessness, and food insecurity.

21
EPIC | Education Policy Innovation Collaborative — Michigan State University

Districts will need sufficient funding to recruit and retain high quality educators to help
students succeed and to provide necessary staff and services to support students in other
ways such as the provision of mental health services. Additional funding for Partnership
districts and low-performing schools and districts generally will be a crucial ingredient in the
recovery efforts. The federal government has taken steps toward improving equitable funding
during the pandemic through the American Rescue Plan Elementary and Secondary School
Emergency Relief (ARP-ESSER) fund. Most of these are formula funds awarded based on Title I
Part A allocations. The state has also prioritized equity by earmarking the non-formula portion
of its ARP-ESSER award as ARP-ESSER equity funds targeting the highest needs districts. Table
2 shows that Partnership districts have been slated for about $700 per student in formula
funds and $100 in ARP-ESSER equity funds, compared with $250 and $20 per student in non-
Partnership districts. The state has also earmarked additional COVID-19 relief funds more
heavily toward Partnership districts through its ARP-ESSER II and III funds.

TABLE 2. ARP-ESSER Funds Per Student


Partnership Non-Partnership
ESSER 1 funds per student $706.88 $252.75
ESSER equity funds per student $107.53 $22.51
ESSER 2 funds per student $2,776.48 $994.64
ESSER 3 funds per student $6,235.57 $2,233.81
ESSER total funds per student
$9,826.46 $3,421.03
(awarded+estimated)
Note: ESSER 1 funds have been awarded. ESSER equity funds are approved amounts. ESSER 2 and 3 are estimates as
of April 20, 2021. Total reflects the sum of all four.

Accountability Policy During and After COVID-19 Will Need to


Consider School Performance Within a Broader Context
Two broad challenges have emerged around school accountability in response to COVID-19.
The first is how to assess schools and districts currently in improvement status given the
pandemic’s effects. The second is how to identify low-performing schools moving forward
while acknowledging the opportunity gap that widened during the pandemic.

With regard to the first challenge, the goals set forth in districts’ Partnership Agreements rely
on valid and reliable test score data from statewide exams (i.e., M-STEP, PSAT, SAT) in addition
to other non-test-score goals. Because the U.S. Department of Education granted waivers from
spring 2020 exams, the state does not have test score data from the 2019-20 school year. While
the Biden administration is requiring that states administer standardized achievement tests in
spring 2021, only students who are learning in-person are expected to take the exams. Because
Partnership districts remain more likely to be teaching students remotely and there is greater
take-up of remote instruction in Partnership districts that give students the option to learn in-

22
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

person or remotely, schoolwide measures of student learning in Partnership districts will be


unable to accurately capture the performance of their students. Moreover, because of the heavy
and inequitable toll the pandemic has taken on students in Partnership districts, it is not apparent
that Partnership districts should be held accountable for their students’ progress in a year when
educators and students have faced enormous challenges and had insufficient opportunity to learn.

Researchers have suggested a number of metrics to unpack differences in test-taking populations


across schools, including examining the percentage of students with comparable test scores
in prior years, comparing students only with similarly performing students in prior years, and
examining best-case academic disparities for students who did not take the exam.18 It remains
to be seen whether these strategies will be feasible given the context described above.

 In terms of the second challenge, identifying the next round of low-performing schools under ESSA
will not occur until the 2022-23 school year using data from the spring 2022 accountability exams.
As long as these exams occur under typical conditions (i.e., in person, during the testing period in
spring 2022, and administered to the same students who took the exams in spring 2019), they can
provide a valid measure of student learning across the three years from spring 2019 to spring 2022.
However, they will not provide a valid measure of school effectiveness because the pandemic’s
effects on schools have varied across communities. Specifically, for accountability systems to use
assessments in a reliable and valid way, all students need the opportunity to learn what will be on
the tests. Students in Partnership communities in particular, who have spent much of the school
year under remote instruction and who are more likely to have experienced personal losses related
to the pandemic, are less likely to have had sufficient opportunity to learn than students in more
affluent communities. Holding districts, schools, and educators accountable for the learning that
occurs during the pandemic without regard to differences in opportunity to learn would penalize
educators in lower income communities for factors outside of their control.

While there is no consensus among education researchers about how states should proceed
with accountability, some have proposed collecting additional data on opportunity to learn and
evaluating test data in the context of this information, and changing the indicators on which
schools will be measured to more accurately reflect education during the pandemic. Finally, even
after in-person schooling and annual accountability testing resumes, accountability policy in
the immediate aftermath of the pandemic that fails to account for inequalities exacerbated by
COVID-19 would misconstrue school effectiveness. Leave-year-out measures of school growth
provide a way to calculate growth from the 2018-19 school year but do not account for disparate
effects of COVID-19 on communities and schools. In the coming years, policymakers will need to
take a comprehensive view of school effectiveness that considers the potential longer term effect
of COVID-19 on student outcomes. Moving forward, some researchers and advocates call for a
more comprehensive approach to assessment that leverages curriculum-embedded formative
assessments and data on student need to create a roadmap for closing opportunity gaps rather
than just identifying achievement gaps.19

23
EPIC | Education Policy Innovation Collaborative — Michigan State University

As Districts Make Plans for the Infusion of COVID-Related Funds,


They Should Consider Ways They Can Leverage Those Funds to
Narrow Opportunity Gaps and Accelerate Learning by Closing
the Digital Divide and Providing Teachers With Professional
Development to Support Students Who Have Experienced Trauma
As we describe above, districts are receiving additional funds to support student learning during
the pandemic, and can invest those funds in resources, structures, and programs that have the
potential to narrow opportunity gaps for their students and build on student experiences to
accelerate learning. Specifically, our analyses find that while children in Partnership communities
are less likely to have access to computers and reliable internet, larger Partnership districts in
particular lacked the resources to provide technology for all students. Investing in technology
that will help close the digital divide can help improve the opportunity to learn for students in
Partnership districts—though technology cannot be a one-time purchase and higher poverty
districts will require additional dollars to refresh technology resources over time. Additionally,
federal education funds are not intended to close the broadband gap. Students need access to
technology devices but their communities need the appropriate infrastructure for fast and reliable
internet connectivity.

In addition to targeting funds toward technology, districts can support students’ socioemotional
learning through a comprehensive system that includes support for socioemotional health as well
as academic growth.20 Our findings suggest that students in Partnership districts in particular have
navigated extraordinary mental health challenges due to the pandemic. Instructional practices
that acknowledge these challenges will be important for educators to address their students’
socioemotional needs. A number of researchers and advocates suggest embedding socioemotional
and trauma-informed practices into student learning.21 Districts could invest relief funds in high
quality professional development for educators to implement trauma-informed practices and
infuse socioemotional learning into their lesson plans and daily routines.22 As educators and
students return to in-person schooling, providing them with the necessary tools to cope with
trauma and build resilience will be critical to accelerate learning and close the opportunity gap for
students in Partnership schools.

24
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

ENDNOTES
1 Hopkins, H., Kilbride, T., & Strunk K.O. (May 3, 2021). 8 Barnum, M. (Feb. 18, 2021). Polls show most—though not
Instructional delivery under Michigan districts’ Extended all parents—are getting the type of instruction they want for
COVID-19 Learning Plans — April update. Education Policy their kids. Chalkbeat. https://www.chalkbeat.org/2021/2/18
Innovation Collaborative. /22289735/parents-polls-schools-opening-remote.

2 Dorn, E., Hancock, B., Sarakatsannis, J., & Viruleg, E. (2020). 9 For districts that indicated 100% of students received
COVID-19 and student learning in the United States: The hurt a single instructional modality, we count their entire
could last a lifetime. McKinsey & Company. enrollment in the selected modality. For districts that planned
to offer families the choice between modes of instruction
3 Karpman, M., Zuckerman, S., & Kenney, G.M. (Oct. 23, 2020). or to provide different subsets of students with different
Uneven recovery leaves many Hispanic, Black and low- instructional modalities, we divide total district enrollment
income adults struggling. Washington, D.C.: Urban Institute. based on the indicated percentage range of students
https://www.urban.org/research/publication/uneven- receiving each modality.
recovery-leaves-many-hispanic-black-and-low-income-
adults-struggling; Gundersen, C., Hake, M., Dewey, A., 10 Kogan, V., & Lavertu, S. (January  27, 2021). The COVID-19
& Engelhard E. (October 2020). Food insecurity during pandemic and student achievement on Ohio’s third-grade
COVID-19. Applied Economic Perspectives and Policy, accessed English language arts assessment. Ohio State University.
January 17, 2021, https://doi.org/10.1002/aepp.13100.
11 Khlaif, Z.N., Salha, S., & Kouraichi, B. (April 27, 2021).
4 Leddy, A.M., Weiser, S.D., Palar, K., & Seligman, H. Emergency remote learning during COVID-19 crisis:
(Nov. 11, 2020). A conceptual model for understanding Students’ engagement. Education and Information
the rapid COVID-19-related increase in food insecurity Technologies, https://doi.org/10.1007/s10639-021-10566-4.
and its impact on health and healthcare. The
American Journal of Clinical Nutrition, 112(5), 1162-1169. 12 Augenblick, Palaich and Associates, & Picus, Odden and
https://doi.org/10.1093/ajcn/nqaa226. Associates. (Jan. 12, 2018). Costing out the resources
needed to meet Michigan’s standards and requirements.
5 Berkowitz, S.A., Cené, C.W., & Chatterjee, A. (Sept. East Lansing, MI: Michigan School Finance Collaborative.
17, 2020). COVID-19 and health equity—time to think https://www.fundmischools.org/wp-content/uploads/2018/
big. New England Journal of Medicine, 383(12), e76. 01/School-Finance-Research-Collaborative-Report.pdf.
https://doi.org/10.1056/NEJMp2021209; Holmes, L.,
Enwere, M., Williams, J., Ogundele, B., Chavan, P., Piccoli, 13 Chingos. M.M., & Blagg, K. (November, 2017). Making sense
T., Chinaka, C., Comeaux, C., Pelaez, L., Okundaye, O., of state school funding policy. Urban Institute.
Stalnaker, L., Kalle, F., Deepika, K., Philipcien, G., Poleon, https://www.urban.org/research/publication/making-sense-
M., Ogungbade, G., Elmi, H., John, V., & Dabney, K.W. state-school-funding-policy
(January 2020). Black-white risk differentials in COVID-19
14 Arsen, D., Delpier, T., & Nagel, J. (January 2019). Michigan
(SARS-COV2) transmission, mortality and case fatality
school finance at the crossroads: A quarter century of state
in the United States: Translational epidemiologic
control. East Lansing, MI: Michigan State University Education
perspective and challenges. International Journal of
Policy Report. http://education.msu.edu/ed-policy-phd/
Environmental Research and Public Health, 17(12), 4322,
pdf/Michigan-School-Finance-at-the-Crossroads-A-
https://doi.org/10.3390/ijerph17124322.
Quarter-Center-of-State-Control.pdf.
6 Greene, S., & McCargo, A. (May 29, 2020). New
15 Kamenetz, A., Trevino, M.A., & Bakeman, J. (Oct. 9,
data suggest COVID-19 is widening housing
2020). Enrollment is dropping in public schools around
disparities by race and income. Urban Institute.
the country. Morning Edition, National Public Radio.
https://www.urban.org/urban-wire/new-data-suggest-
https://www.npr.org/2020/10/09/920316481/enrollment-
covid-19-widening-housing-disparities-race-and-income;
is-dropping-in-public-schools-around-the-country;
Karpman, M., Zuckerman, S., & Kenney, G. (Oct. 27, 2020).
Lee, V., Gutierrez, E., & Blagg, K. (Oct. 6, 2020). Declining
Uneven recovery leaves many Hispanic, Black, and low-
school enrollment spells trouble for education funding. Urban
income adults struggling. Urban Institute.
Institute (blog), https://www.urban.org/urban-wire/declining-
7 Karpman et al. (October 2020). school-enrollment-spells-trouble-education-funding.

25
EPIC | Education Policy Innovation Collaborative — Michigan State University

ENDNOTES
16 Guerrant, K.L. (Aug. 21, 2020). Memorandum to local and 19 Darling-Hammond, L., Schachner, A., Edgerton, A.,
intermediate school district superintendents public school Badrinarayan, A., Cardichon, J., Cookson, P.W., Griffith,
academy directors: Return to learn law details. Lansing, MI: M., Klevan, S., Maier, A., Martinez, M., Melnick, H.,
Michigan Department of Education. Truong, N., & Wojcikiewicz, S. (August 2020). Restarting
and reinventing school: Learning in the time of COVID
17 Strunk, K.O., Burns, J., Torres, C., Mcilwain, A., Waldron, S.W., and beyond. Palo Alto, CA: Learning Policy Institute,
& Harbatkin, E. (October 2020). Partnership turnaround: http://learningpolicyinstitute.org/product/restarting-
Year two report. Education Policy Innovation Collaborative. reinventing-school-covid.
https://epicedpolicy.org/partnership-turnaround-
year-two-report/. 20 Darling-Hammond et al. (August 2020).

18 Ho, A. (Feb. 26, 2021). Three test-score metrics that 21 Darling-Hammond, L., & Hyler, M.E. (Aug. 7, 2020).
all states should report in the COVID-19-affected Preparing educators for the time of COVID...and beyond.
spring of 2021. Harvard Graduate School of Education, European Journal of Teacher Education, 43(4), 457-65.
https://scholar.harvard.edu/files/andrewho/files/ https://doi.org/10.1080/02619768.2020.1816961.
threemetrics.pdf.
22 Srinivasan, M. (June 21, 2019). Three keys to infusing
SEL into what you already teach. Greater Good Magazine.
https://greatergood.berkeley.edu/article/item/three_
keys_to_infusing_sel_into_what_you_already_teach.

26
COVID-19 and Michigan’s Lowest Performing Schools | May 2021

DATA APPENDIX
This data appendix provides more detail on the data sources provided in Table 1 of the brief as well as the measures we drew from
each of these data sources.

U.S. Census Bureau American Community Survey (ACS). We drew from the ACS for data related to county characteristics
that might exacerbate the negative health, economic, and education-related effects of COVID-19. Most of these measures
came from the 2018 five-year ACS estimates, the most recent numbers that cover all counties in the state for the variables
of interest (See: the 2014-2018 5-Year Estimates, available at: https://www.census.gov/programs-surveys/acs/technical-
documentation/table-and-geography-changes/2018/5-year.html). The estimate for median household income is from the 2019
1-year estimates. By using measures from data collected prior to the pandemic, we are able to understand baseline differences in
Partnership and non-Partnership communities. We present ACS data on race, ethnicity, poverty, educational attainment, internet
access, computer and device access, and receipt of Supplemental Nutrition Assistance Program (SNAP) benefits. Federal SNAP
benefits, colloquially known as food stamps, provide assistance for low-income families to purchase food. A family of four in Michigan
is eligible for SNAP if its household income is $34,060 or less.

District revenue data. The Michigan Department of Education (MDE) and the Center for Educational Performance and
Information (CEPI) provide district-level data on revenues and revenue sources. We used district revenue data from the
2019-20 school year. We restricted the data to local school districts only and classified revenues as federal, state, local, or
other according to revenue major class codes as described in the Michigan Public School Accounting Manual (available at
https://www.michigan.gov/mde/0,4615,7-140-6605-21321--,00.html).

Continuity of Learning (COL) plans. To examine the strategies districts put into place for educating their students during the pandemic,
we drew from COL plans that all districts submitted to the state after the state suspended in-person instruction in the 2019-20 school
year. These documents, created in April 2020 to comply with Executive Order 2020-35, outline district plans for providing instruction,
monitoring student learning, and supporting students during the pandemic. (See EPIC’s policy brief on the COL plans, available at
https://epicedpolicy.org/how-did-michigan-school-districts-plan-to-educate-students-during-covid-19/, for more information on
the COL plans). We drew a subset of items from the COL plans that align with the community characteristics illuminated by the
ACS data. Specifically, we identified items related to meal distribution plans, technology provision (plans to provide students with
electronic devices, Wi-Fi hotspots, or resources about internet access), and academic or socioemotional supports (plans to check in
with students, provide mental health resources, make referrals to outside support services, and provide access to counselors or mental
health professionals).

Reconfirmed Extended COVID-19 Learning (ECOL) Plan Monthly Questionnaire. To assess the extent to which Partnership districts
are educating students in-person and remotely, we used the state’s ECOL Plans Monthly Questionnaire, through which districts
submit their plans for instructional modality for each month (See EPIC’s monthly analyses of instructional modality plans, available at
https://epicedpolicy.org/ecol-reports/, for more information on the Reconfirmed COVID-19 Learning Plan Monthly Questionnaires).
The questionnaire asks districts what percentage of students they plan to instruct fully in person, fully remote, or in hybrid format
for each month. For each district in each month from September 2020 through February 2021, we collapsed these responses into
three mutually exclusive categories. The first category, “in-person option,” classifies districts that have any students attending fully in
person. The second, “hybrid,” classifies districts without a fully in-person option that have any students attending a hybrid model. The
final category, “fully remote,” identifies districts in which all students attend remotely.

COVID-19 case, death, and positivity rates. We retrieved publicly available COVID-19 confirmed case counts, death counts, and test
positivity rates from the Michigan Department of Health and Human Services (MDHHS). We converted case and death counts to
rates per 100,000 population using the 2019 annual county population estimates from the U.S. Census Bureau. For all three measures,
we then calculated 7-day rolling averages for each day to account for reporting idiosyncrasies—especially in smaller counties and on
weekends and holidays. The state reports cases and deaths beginning March 1, 2020. We, therefore, constructed these 7-day averages
beginning March 6. The first available case positivity rates vary by county, with the first county reporting positivity rates beginning
March 20, 2020 and the last reporting March 29, 2020. Most counties began reporting positivity rates between March 16 and March
19, 2020. We report 7-day positivity averages beginning April 1. For each county, we also calculated the cumulative confirmed case and
death rates per 100,000 as a measure of the cumulative toll to date.

For student-weighted figures, we used enrollment data from the Center for Educational Performance and Information, Student Count
Report, available at https://www.mischooldata.org/DistrictSchoolProfiles2/StudentInformation/StudentCounts/StudentCount2.aspx.

27
EPIC

Education Policy Innovation Collaborative


MICHIGAN STATE UNIVERSITY (517) 884-0377
236 Erickson Hall | 620 Farm Lane EPICedpolicy@msu.edu
East Lansing, MI 48824 www.EPICedpolicy.msu.edu

RESEARCH WITH CONSEQUENCE

You might also like