Michigan’s K-12 Virtual Learning Effectiveness Report, 2019-20

Table of Contents 

Past Effectiveness Reports 

View Infographic Download PDF Version Watch Report Overview

Introduction

This report presents analysis of information on virtual learners reported by schools to the state and shares findings in a highly consumable way to aid the evaluation of virtual learning programs. This year’s report is the eighth edition of this annual publication and completes 10 years of data on K-12 virtual learning in Michigan. Past reports are available through the Michigan Virtual website.

The report is organized into several sections. The first section looks at schools as the unit of analysis. The next section focuses on the virtual courses taken. The third section focuses on students. There is also a brief section containing maps of virtual use. Each section is meant to capture the essential findings without being overly data intensive; however, data tables have been included in the appendices to provide those interested with more in-depth information. Information about the report’s methodology is captured in Appendix A. Please note that in some tables and figures, the percentage data may not sum to 100% due to rounding.

Schools

Fast Facts

• 613 school districts reported at least one virtual enrollment. This represented 68% of Michigan school districts.
• 57% of the 1,225 schools with virtual enrollments had 100 or more virtual enrollments.
• 75% of schools with virtual enrollments had a general education school emphasis; 24% had an alternative education emphasis.
• 87% of schools with virtual learning were LEA schools.
• LEA schools accounted for 61% of the virtual enrollments; PSA schools generated 37% of the virtual enrollments.
• 54% of virtual enrollments came from schools with part-time virtual learning options.
• LEA schools had the most full-time virtual schools (58).
• 98% of virtual enrollments came from schools with 100 or more virtual enrollments.
• About 81% of virtual enrollments came from students in grades 9-12.
• 34% of virtual enrollments came from suburban schools, the most of any locale.
• Schools with a general education emphasis had a 68% virtual pass rate, outperforming those with an alternative education emphasis, which had a pass rate of 43%.
• 28% of schools had a school-wide virtual pass rate of 90% to 100%.

Number of Districts and Schools

For the 2019-20 school year, 613 districts reported having at least one virtual enrollment. This represented 68% of the 901 Michigan public school districts for the year. See the MI School Data Report for a breakdown of the district count.

Within those districts, 1,225 schools reported virtual enrollments, the same number as the prior year. Of the 1,225 schools reporting virtual enrollments in 2018-2019 (the prior year), 145 of them (12%) did not report any virtual enrollments in 2019-2020 (this year). Those schools had accounted for a total of 18,333 virtual enrollments in 2018-2019. There were also 145 schools that reported virtual enrollments in 2019-2020 but had not reported any the prior year. These schools added 13,430 virtual enrollments this year.

That means 88% (1,080) of schools in the dataset reported virtual enrollments in both 2018-19 and 2019-2020. A total of 554 schools reported increased virtual enrollments (125,574 more) whereas 519 schools reported decreases in virtual enrollments (87,119 fewer). Seven schools reported the same number both years. Taken together, the annual growth in the number of virtual enrollments is due to increased enrollments from schools with pre-existing virtual programs rather than the addition of new schools with virtual learning programs.

By Grade Level

Across the 1,225 schools, 672,682 virtual enrollments were taken. Students in 12th grade enrolled in the most virtual courses (190,850), representing approximately 28% of all virtual enrollments. The overall pass rate for virtual enrollments was 56%, a one percent increase over the prior three years. See Table F1 for a more specific breakdown of all the completion statuses. This ranged from a high of 65% in 4th grade to a low of 42% in 9th grade. See Table B1 for more information. Consistent with findings from previous years, virtual learners passed their virtual courses at a lower rate (56%) than they passed their non-virtual courses (78%). This gap is approximately the same as the prior year. See Table B2.

By School-Level Virtual Pass Rate

Of the 1,225 schools with virtual enrollments, 341 or 28% had school-level virtual pass rates of 90% to 100%. More than half of schools (56%) had virtual pass rates of 70% or better. See Table B3.

By Entity Type

LEA schools (61%) and PSA schools (37%) accounted for almost all the virtual enrollments. Virtual enrollments came from 1,068 (87%) LEA schools while only 116 (9%) of the schools were PSAs. See Table B4. LEA schools had a higher pass rate (57%) than PSA schools (53%). See Table B5 or, for a more in-depth look at the completion statuses, see Table F2.

By Full-Time Virtual Schools

The number of full-time virtual schools (77) dropped by two from the prior year. Fifty-eight of the 77 full-time virtual schools (75%) were LEA schools. PSA schools (16) accounted for 21% of the full-time virtual schools. See Table B6. Forty-two percent of students attending full-time virtual schools did so at LEA schools with 58% attending a PSA school. Unlike the prior year, there was a sizable performance difference between these two types. LEA schools had a virtual pass rate of 42% (-9% from the prior year), and PSAs had a virtual pass rate of 58% (+9% from the prior year). See Table B7 and Table F3. In total, 46% of virtual enrollments came from cyber or full-time virtual schools – the same percentage as the prior year.

A quick note about full-time virtual schools: Historically, full-time virtual schools have only provided students with 100% of their learning online. Thus, it was safe to designate all enrollments from such a school as being part of a full-time virtual program. Over the last few years, however, LEAs have started to add full-time virtual options to their offerings. In some cases, this is as a separate school, which makes it analogous to the cyber schools. However, increasingly, it seems that schools are offering multiple forms of online learning from the same building code. This can be seen in the Education Entity Master where schools can report educational settings including “Full Virtual,” “Face Virtual,” and “Supplemental Virtual.” See page 16 of the Educational Entity Master Glossary for more information on these field values. This means that some schools report various forms of virtual learning from a single building code. In the 2019-20 dataset, 98% of the virtual enrollments reported by schools with the full-time virtual designation came from students who only had virtual enrollments reported by the school. This rate will be something worth tracking in subsequent years to see if that figure moves demonstrably.

By Part-Time Virtual Schools

About 94% of the schools offering virtual learning do so to supplement their face-to-face course offerings. These 1,148 schools, referred to in this report as part-time virtual schools, were predominantly LEA schools (88%). See Table B8. Eighty-seven percent of the part-time virtual students were enrolled through LEA schools and 11% through PSA schools. LEA schools accounted for almost 295,000 virtual enrollments or 80% of the part-time enrollments. In total, part-time virtual enrollments accounted for 54% of all the virtual enrollments for the year. LEA schools had a pass rate of 63% whereas PSA schools had a pass rate of 38%. Overall, the pass rate for the part-time virtual schools (59%) was over seven percentage points higher than the rate for the full-time virtual schools (52%). See Table B9 and Table F4.

By School Emphasis

Three out of four schools with virtual learning were designated as General Education and produced 349,674 (52%) of the virtual enrollments. Schools with Alternative Education as their emphasis accounted for 317,720 (47%) of the virtual enrollments. See Table B10. There was a considerable difference in virtual pass rates between these two types of schools. General Education schools had a 68% virtual pass rate, whereas Alternative Education schools had a 43% virtual pass rate (see Table B11 and Table F5), though this varied by entity type. LEA schools, for instance, had a 73% virtual pass rate for General Education schools and a 45% virtual pass rate for Alternative Education schools. See Table B12.

By Number of Virtual Enrollments

Over half of schools with virtual enrollments (57%) had 100 or more virtual enrollments. These schools were responsible for 98% of the virtual enrollments (656,837). As has been observed in previous years, schools with less than 10 virtual enrollments were the next highest percentage of schools with 14%; however, they generated less than .1% of the virtual enrollments. See Table B13.

Another trend that continued was that, in general, schools with fewer virtual enrollments per student performed better. Consider, for instance, that 37% of schools with an average of one to two virtual enrollments per virtual learner had a virtual pass rate of 90% to 100%, whereas only 16% of schools with an average of four or more virtual courses per virtual learner had a 90% to 100% pass rate. See Table B14.

By Locale

Rural schools represented about 36% of schools with virtual enrollments. Suburban settings provided the second most schools with 31%. Suburban schools, however, tallied the largest percentage of the virtual enrollments at 34%. All four locales (urban, suburban, town, and rural) had more than 100,000 virtual enrollments. See Table B15. In each of the four locales, schools with 100 or more virtual enrollments accounted for the largest percentage of schools. Similarly, schools with less than 25 virtual enrollments was the second most likely scenario. See Table B16. Virtual pass rates varied by locale with town schools having the highest virtual pass rate at 60% and city having the lowest at 47%. Both city schools (11%) and those not specified (14%) had the highest percentage of schools with pass rates less than 20%. See Tables B17 and B18.

By School Poverty Levels

The overall level of poverty for a school yielded insightful analyses. Schools were categorized into one of four categories based on the percentage of all learners at the school (not just virtual learners) that were in poverty:

• Low Poverty (<=25%)
• Mid-Low Poverty (>25% to <=50%)
• Mid-High Poverty (>50% to <=75%)
• High Poverty (>75%)

In comparison to all schools in Michigan in each category, the percentage of schools with virtual learners was consistent across Low (38%), Mid-Low (39%), and Mid-High (38%) Poverty schools. In contrast, only about 27% of the state’s High Poverty schools had virtual learners. See Table B19.

While High Poverty schools represented only 20% of schools with virtual programs (241), they accounted for 39% of the virtual enrollments. Mid-High Poverty schools accounted for 36% of the enrollments. The pass virtual pass rate for Low Poverty schools was 66% compared to 50% for High Poverty schools. See Table B20.

Courses

Fast Facts

• 672,682 virtual enrollments were taken by Michigan K-12 students; the overall pass rate for virtual enrollments was 56%.
• Virtual enrollments were spread across 964 different course titles.
• 67% of virtual enrollments occurred in the core subject areas of English Language and Literature, Mathematics, Life and Physical Sciences, and Social Sciences and History.
• The course titles with the highest enrollments for each core subject were:
  • English Language and Literature: English 9, English 10, English 11, and English 12
  • Mathematics: Geometry, Algebra II, Algebra I, and Consumer Math
  • Life and Physical Sciences: Biology, Chemistry, Earth Science, and Physical Science
  • Social Sciences and History: U.S. History—Comprehensive, Economics, World History—Overview, and World History and Geography
• The virtual pass rates for each core subject were:
  • English Language and Literature: 54%
  • Mathematics: 52%
  • Life and Physical Sciences: 54%
  • Social Sciences and History: 58%
• 31 different Advanced Placement (AP) courses were taken virtually.
• The percentage of enrollments was fairly consistent by subject area across rural, town, suburban, and city schools.
• Online courses (defined as including a teacher in the virtual environment) produced 83% of the virtual enrollments. Digital learning (without a teacher in the virtual environment) and blended learning (some virtual, some face-to face instruction) each accounted for 8% of the virtual enrollments.

Number of Courses

The 672.682 virtual enrollments came from 964 different course titles, as determined by unique SCED codes.

Courses by Subject Area

English Language and Literature was the subject area with the highest virtual enrollment with 129,014 enrollments – 19% of all virtual enrollments. Mathematics, Social Sciences and History, and Life and Physical Sciences were the other subject areas with 10% or more of the virtual enrollments. In high enrollment subject areas (greater than 25,000 virtual enrollments), virtual pass rates varied from a low of 51% in Miscellaneous to a high of 58% for Social Studies and History as well as Fine and Performing Arts. See Table C1 and Table F6. The virtual pass rates were consistently lower than the non-virtual pass rate for the virtual learners in their non-virtual courses, a trend observed in past years. Exceptions to this were Nonsubject Specific and Architecture and Construction. See Table C2.

Highest Virtual Enrollment Courses

For English Language and Literature, the most highly enrolled in virtual courses were 9th, 10th, 11th, and 12th grade English/Language Arts. Of those four, the pass rate was lowest for 9th grade English/Language Arts (41%) and consistently rose for each subsequent grade level to finish at 62% for 12th grade English/Language Arts. See Table C3.

In Mathematics, Geometry, Algebra II, and Algebra I had the most enrollments. The pass rate across the top 10 most enrolled-in virtual mathematics courses ranged from a low of 39% for Algebra I, Algebra I – Part 1, and Pre-Algebra to a high of 68% for Consumer Math. See Table C4.

Biology, Chemistry, and Earth Science were the three course titles responsible for more than 10% of the virtual enrollments in Life and Physical Sciences courses. A quarter of all Life and Physical Sciences virtual courses were in Biology. Life and Physical Sciences-Other had the lowest pass rate (39%) of those in the top 10; the highest was Environmental Science at 60%. See Table C5.

For Social Sciences and History, the four course titles of U.S. History–Comprehensive, Economics, World History–Overview, and World History and Geography each yielded more than 10% of the virtual enrollments. Pass rates for the top 10 most enrolled in courses ranged from a low of 46% in U.S. History-Other to a high of 67% for Sociology. See Table C6.

Thirty-one AP courses were taken virtually in 2019-20. AP Psychology was the most popular course accounting for 19% of the 3,778 AP enrollments. The pass rate for AP courses taken virtually was 86%. See Table C7. The pass rate for non-virtual AP courses taken by virtual learners was 94%.

Subject Area Enrollments by Locale

Course enrollment patterns were consistent across locales. For instance, Mathematics represented between 16% and 19% of the virtual enrollments for all (rural, town, suburban, city, and not specified) locales. The range was 2% (13% to 15%) for Life and Physical Sciences and 7% (14% to 21%) in English Language and Literature. See Table C8. Pass rates in virtual courses also varied across subject areas and locale. For instance, in English Language and Literature, pass rates fell between 44% for city schools to 59% for suburban schools. In Mathematics, pass rates ranged from 44% (city) to 54% (suburban and town). See Table C9.

Subject Area Enrollments by Student Sex

Males and females enrolled in subject areas in similar proportions. In the four highest enrollment subject areas (English Language and Literature, Mathematics, Life and Physical Sciences, and Social Sciences and History), the proportion of enrollment from males and females was within one percent of each other. Pass rates did, however, show more variability by student sex; in most cases, females outperformed males – a trend that has been consistent with past years. Males had a 54% pass rate whereas females had a 58% pass rate. See Table C10.

Courses by Virtual Method

Schools classified the virtual courses into one of three methods: Blended Learning, Digital Learning, or Online Learning. See pages 356 and 357 of the Michigan Student Data System Collection Details Manual Version 4.0.

• Blended Learning – A hybrid instructional delivery model where pupils are provided content, instruction, and assessment at a supervised educational facility where the pupil and teacher are in the same physical location and in part through internet-connected learning environments with some degree of pupil control over time, location, and pace of instruction. For a course to be considered blended, at least 30% of the course content is delivered online.
• Digital Learning – A course of study that is capable of generating a credit or a grade that is provided in an interactive internet-connected learning environment that does not contain an instructor within the online environment itself. There may be a teacher of record assigned to the course, but this teacher does not provide instruction to students through the online environment. For a course to be considered online as opposed to blended, all (or almost all) the course content is delivered online.
• Online Course – A course of study that is capable of generating a credit or a grade that is provided in an interactive internet-connected learning environment, where pupils are separated from their teachers by time or location, or both. For a course to be considered online as opposed to blended, all (or almost all) the course content is delivered online.

Blended Learning enrollments accounted for 8% of the virtual enrollments and had a pass rate of 53%. Digital Learning also totaled 8% of the enrollments with a 52% pass rate. Online courses represented most of the enrollments (83%) and yielded a pass rate of 56%. See Table C11.

Students

Fast Facts

• 121,900 K-12 students took at least one virtual course which represented 8% of Michigan public school students.
• 86% of virtual learners were in high school; 32% were seniors and 21% were juniors.
• 47% of virtual learners passed all their virtual courses. Twenty-one percent of virtual learners did not pass any of their virtual courses.
• Of the 25,780 students who did not pass any of their virtual courses, 41% took only one or two courses. Over 11,900 students took and did not pass five or more virtual courses with 3,228 students taking and not passing 11 or more virtual courses.
• Female students had a higher pass rate (58%) than did males (54%).
• Students in poverty made up the majority of virtual learners (59%) and virtual enrollments (68%). Students in poverty also had a lower pass rate (50% v. 68%).
• Part-time virtual learners had higher pass rates (59%) compared to full-time virtual learners (52%). This also included students in poverty (51% v. 49%).
• Students using special education services made up 12% of the virtual learners.
• Pass rates were higher for students taking fewer virtual courses. Students taking one or two virtual courses had a 76% pass rate compared to 52% for those taking five or more.
• White students represented 67% of virtual students; African-Americans were 17%.

By Grade Level

For the 2019-20 school year, 121,900 Michigan K-12 students, approximately 8% of students in the state, took at least one virtual course. This was a 1% increase in the number of virtual learners compared to what was reported for 2018-19. Only about 6% of the state’s virtual learners were in grades K-5. Grades 6-8 were responsible for about 9% of the virtual learners. High school grade levels generated 86% of the virtual learners. About 32% of virtual learners were high school seniors and 21% were juniors. See Table D1.

By Student Sex

There were slightly more females (61,549) enrolled in virtual courses than males (60,367), though from a percentage perspective, each represented about half of the population. Females had a 4% higher pass rate (58% compared to males at 54%), continuing the trend seen in past years of females outperforming their male counterparts on this measure. See Table D2 and Table F7.

By Race/Ethnicity

White students made up 67% of virtual students with African American students totaling the second highest percentage with 17%. Asian students had the only pass rate (74%) above 60%. See Table D3 and Table F8.

By Poverty Status

Fifty-nine percent of virtual learners were classified as living in poverty. This is about 8% higher than the percentage of 2019-20 K-12 students statewide who were economically disadvantaged. See Student Enrollment Count Report. Students living in poverty took 68% of the virtual enrollments for the year. The pass rate for students in poverty (50%) was 18 percentage points lower than students who were not in poverty (68%). See Table D4 and Table F9.

In addition to the performance gap between those in poverty and those not in poverty, there were also differences in non-virtual pass rates. Virtual learners in poverty had a 70% pass rate in their non-virtual courses, 20 percentage points better than their virtual pass rate. Interestingly, students not in poverty had an 86% pass rate in their non-virtual courses, an improvement of 19 percentage points over their virtual pass rate. Thus, students in poverty had a slightly larger performance gap between their virtual and non-virtual pass rates than did students who were not in poverty. See Table D5.

Differences were apparent by virtual type. Sixty-nine percent of full-time virtual learners were in poverty compared to 55% of part-time virtual learners. The pass rate for full-time virtual learners in poverty was 49% compared to 51% for part-time virtual learners. See Table D6.

About 6% of all Michigan K-12 students who attend Low Poverty schools were virtual learners. The percentage steadily climbed as the school’s poverty level increased. Seven percent of the state’s students in Mid-Low Poverty schools were virtual learners, and it was 10% for Mid-High. Eleven percent of students in High Poverty schools were taking courses online in the 2019-2020 school year. See Table D7.

By Special Education Status

Students using special education services made up 12% of the virtual learners and 13% of the virtual enrollments. These percentages are somewhat similar to the statewide percentage of students using special education services (13.41%) for the 2019-20 school year. See the Student Enrollment Counts Report. Students using special education services had a virtual pass rate of 50% compared to 57% for those who did not. See Table D8 and Table F10.

By Full-Time or Part-Time

A little more than a quarter of students (32,976) were enrolled in cyber or full-time virtual schools. Students in these schools accounted for 306,495 or 46% of the virtual enrollments for the year. The pass rate for full-time virtual students was 52%. Three out of four virtual learning students were part-time virtual learners, taking some courses virtually to supplement their face-to-face schedule. This subset made up 54% of the virtual enrollments and had a pass rate of 59%. See Table D9. This rate is about 19% lower than their pass rate in face-to-face courses (78%). See Table D10.

By Non-Virtual Course Performance

Part-time virtual learners with at least three non-virtual courses were classified into one of three categories based on their success in those non-virtual courses. The three categories were:

• Passed all Non-Virtual Courses
• Did Not Pass 1 or 2 Non-Virtual Courses
• Did Not Pass 3 or More Non-Virtual Courses

In total, 61% of virtual learners had at least three or more non-virtual enrollments. Of that group, 48% of students passed all their non-virtual courses, 20% did not pass one or two, and 32% did not pass three or more. There were clear differences in virtual pass rates between the three categories. Students passing all of their non-virtual courses had an 82% virtual pass rate. Students who did not pass one or two non-virtual courses had a virtual pass rate of 67%, and those with the lowest non-virtual success had a virtual pass rate of only 43%. See Table D11.

By Virtual Course Performance

Forty-seven percent of virtual learners passed every virtual enrollment they took. Twenty-one percent did not pass any of their virtual enrollments, and 32% passed some, but not all of their virtual enrollments. Students who passed all of their virtual courses were responsible for 30% of the virtual enrollments. Students with mixed success generated 52% of the enrollments, and students who did not pass any of their virtual courses contributed 19% of the virtual enrollments. These statistics are fairly similar to last year. See Table D12.

For the students who did not pass any of their virtual courses, 41% only took one or two virtual courses. On the other hand, 11,932 students did not pass five or more virtual courses, and 3,228 students did not pass 11 or more virtual courses. Further analysis of students failing all of their 11 or more virtual courses showed 87% of these students had a single school report data for them. Roughly half of these students came from part-time virtual programs. Over 540 students were using special education services (17%) and 2,728 of these students (85%) were in poverty. See Table D13 and Table F11.

What Table F11 makes clear is that for students who do not pass any of their virtual enrollments, “withdrawns” and “incompletes” were rampant. For the virtual enrollments from students who did not pass any of their virtual enrollments, 41% had a “Withdrawn” status (exited, failing, or passing) and another 21% were classified as “Incomplete.” For those taking 11 or more virtual courses, 35% had a “Withdrawn” status and 20% were marked “Incomplete.” In each case, only 27% and 32% of the virtual enrollments, respectively, were actually classified as “Completed/Failed.” Please see the section on Pass Rate Calculations for more elaboration on the impact of such issues on pass rates.

By Virtual Usage

Generally speaking, virtual learners did better when they took fewer virtual courses. Students taking one to two virtual courses had a pass rate of 76% compared to a pass rate of 63% for those taking three to four virtual courses and a pass rate of 52% for students taking five or more virtual courses. About 42% of students fell under the description of taking one or two virtual courses; however, 44% were found to have taken five or more virtual courses during the year. See Table D14.

State Assessment

Fast Facts

• State assessments for 2019-2020 were suspended due to the COVID-19 pandemic.

Annual state assessments were not administered for the 2019-2020 school year due to the COVID-19 pandemic closures. As a result, no data exists to conduct such analyses for this year.

Maps

Alpena-Montmorency-Alcona, Berrien, COOR, Gogebic-Ontonagon, Mecosta-Osceola, and St. Joseph ISDs/RESAs had at least 15% of students in their service areas take a virtual course in 2019-20. In total, there were 18 ISDs/RESA with 10% or more of the students taking virtual courses. See Figure 2.

  Figure 2. 2019-20 Percentage of Students Who Took a Virtual Course (Non-Cyber) by ISD

One in five students attending a PSA cyber school resided within the Wayne RESA service area. The Macomb and Oakland ISD service area were the only other ISDs with 1,000 or more students attending PSA cyber schools. Genesee, Kent, Ingham, and St. Clair County ISDs/RESAs each had between 500 and less than 1,000 students from their area attending PSA cyber schools. See Figure 3.

Figure 3. 2019-20 Count of PSA Cyber School Students by Resident ISD

Conclusion

This year’s report represents the 10th year of data on the effectiveness of virtual learning in Michigan’s K-12 system. Many trends witnessed in past years continue to exist. See Table 1. The use of virtual learning as evidenced by the number of virtual learners and virtual enrollments continued to grow. The number of schools remained constant, but the virtual pass rate rose for the first time in the last 4 years.

The overall pass rate remains a reason for pessimism. Successful implementations of virtual learning are outnumbered by poorly performing programs, and there are too many students having little to no success. Almost a quarter of students did not pass any of their virtual courses with over 11,900 students taking at least five virtual courses and passing none of them. These practices need to stop. They need to be replaced by the practices at the more than 540 Michigan schools with virtual pass rates of 80% or higher. In these higher performing schools, the data show:

• Successful virtual programs can support various numbers of students, enrollments, and courses offerings – These schools showed success with less than 10 students (32%) and more than 50 students (38%). Some offered few enrollments (over 100 schools had less than 10 virtual enrollments) while others offered many (over 200 schools had 100 or more). They also varied in the numbers of course titles offered. About 40% offered 10 or fewer virtual courses titles. About one in five had between 26 and 50 courses, and 14% of the schools had students in more than 50 different virtual courses.
• LEA and PSA schools can offer successful virtual programs – Over 45 percent of LEA schools with virtual programs had schoolwide virtual pass rates of 80% or higher. For PSA schools, it was about one out of every three. Both traditional school districts and charter districts can run successful virtual programs.
• Schools in cities, suburbs, towns, and rural settings are proving virtual learning success – There were more than 80 city and 80 town schools with virtual pass rates of 80% or higher. There were over 175 for suburban locales, and 199 of such schools in rural settings. These schools are proving virtual learning can succeed across the various geographies of the state.
• These schools show strong results across students of different race/ethnicities – Whereas a sizable performance gap exists for students based on their race/ethnicity when looking at all virtual programs, those gaps close substantially for these 540+ schools. Consider that the pass rate for African-American or Black students was 86% in these schools (compared to 45% in the overall virtual population) and the gap between it and the White virtual pass rate in these schools dropped to 4% compared to 15% in the overall virtual population. For each known race/ethnicity, the virtual pass rates were above 80%.
• Students in poverty are succeeding in these virtual programs – Recall that across the entire state, students in poverty had a pass rate (50%) that was 18 percentage points lower than those virtual students who were not in poverty. For these 540+ schools, the virtual pass rate for students in poverty rose to 86% — much closer to the 92% virtual pass rate for the students in those schools who were not in poverty. Students in poverty continued to represent a large percentage of virtual learners (42%) and virtual enrollments (49%) in these schools, but quite a bit smaller than the 59% of virtual learners and 68% of virtual enrollments seen across all virtual programs across the state. Additionally, a school’s overall poverty level was a strong predictor of virtual program success. Sixty-six percent of Low Poverty schools with virtual learners achieved virtual pass rates of 80% or higher virtual. It was 54% of the Mid-Low Poverty schools, 37% of Mid-High Poverty schools, and only 27% of High Poverty schools. While these 64 High Poverty schools showed it is possible, it is considerably rarer than it is for Low Poverty schools.
• Both full- and part-time programs can run effective virtual programs, but success is rarer for full-time programs – These 540+ schools show that 47% of part-time programs were able to yield schoolwide virtual pass rates of 80% or higher. It was more difficult for full-time programs to achieve similar success. Only nine of the 77 full-time programs (12%) reached the 80% pass rate mark.
• Both general education and alternative education programs reached 80% school-wide virtual pass rates – There were 494 general education schools in Michigan that achieved schoolwide virtual pass rates of 80% or higher. These 494 schools represented 54% of general education schools with virtual programs. For alternative programs, 45 schools reached this mark. As a percentage of alternative programs, it represented just 16% of such schools, indicating that while possible, this threshold of success remains a sizable challenge.
• Virtual students can perform at or above their face-to-face performance level – In these 540+ schools, there were 7,500 virtual learners who took a minimum of three virtual courses and had data for a minimum of three non-virtual courses. Eighty percent of these students had virtual pass rates that met or exceeded their non-virtual pass rates.

Clearly, these schools support the conclusion that online learning can and does work for many schools and students. We need more of these kinds of programs.

The data in this report can help schools and stakeholders evaluate their virtual learning programs. It can yield important insight and inform critical conversations about what is working and for whom it is working, and what is not working and under what circumstances those results are occurring. Michigan Virtual has created many free resources that can assist schools in reflecting upon and improving their virtual programs. These resources include an expanding series of practical guides designed for students, parents, teachers, mentors, school administrators, and school board members. Michigan Virtual also provides quality reviews of their online learning programs to Michigan schools at no cost.

Appendix A – Methodology

COVID-19 Impact

Readers should note that the COVID-19 pandemic significantly disrupted school in the spring of this school year and likely impacted the data from the school year.

About the Data

The data for this report came from the following sources:

• Michigan Student Data System – School Year 2019-2020;
• Educational Entity Master (EEM);
• Michigan Student Data System Teacher Student Data Link (TSDL) – Collection Year 2019-2020; and
• Michigan’s K-12 Virtual Learning Effectiveness Report, 2018-19 – Used for comparing this year’s data with the 2018-19 school year.

Because the data for this report incorporates this variety of sources, the findings within may differ from those found through the MI School Data portal which may use different query parameters.

The majority of enrollments classified as virtual in this report were treated as such due to the TSDL virtual method field indicating virtual delivery. However, this field is known to contain inaccuracies. For the purposes of this report, additional methods were used to identify enrollments with a high likelihood of having been delivered virtually. Each of the methods used, along with the percentage of enrollments it contributed to the total, are outlined below.

• TSDL Virtual Method Flag = Yes. Enrollments where the TSDL virtual method field was set to “Blended Learning,” “Digital Learning,” or “Online Course” were treated as virtual. According to the TSDL Data Collection Manual, the virtual method field “indicates the type of virtual instruction the student is receiving. This could be virtual learning, online learning or computer courses; distance learning; or self-scheduled virtual learning” (see page 356). This strategy yielded 99% (664,380) of the virtual enrollments.
• TSDL Local Course Title Field References Michigan Virtual. The strategy of searching the local course title field for common references to Michigan Virtual yielded less than 1% (1,112) of the virtual enrollments. The wild card search criteria for Michigan Virtual consisted of the following: ‘%MI Virtual%’, ‘%Mich Virt%’, ‘%MIVHS%’, ‘%MIVS%’, ‘%MVS%’, ‘%MVU%’, ‘%VH’, ‘%MVHS%’, ‘%MIVU%’, ‘%VHS%’, ‘MV%’, ‘%MV’, ‘%Michigan Virtual%’, ‘%IS: MV%’, ‘%IS:MV%’, ‘%MI Virt%’, ‘%MV HS%’, ‘Virtual HS%’, and ‘Mich. Virtual High School%’.
• Local Course Title Field References Common Third Party Providers. Searching the local course title field for common references to known third-party providers of virtual courses yielded less than 1% (2,331) of the virtual enrollments. The wild card search criteria for common third-party providers consisted of the following: ‘%Apex%’, ‘APX%’, ‘%Aventa%’, ‘%BYU%’, ‘%Brigham%’, ‘%Compass%’, ‘%Edgen%’, ‘%2020%’, ‘%20/20%’, ‘%20-20%’, ‘%E20%’, ‘%Edison%’, ‘%FLVS%’, ‘%FVS%’, ‘%GenNet%’, ‘%Gen Net%’, ‘%K12 Virtual%’, ‘%K12:%’, ‘%K12vs%’, ‘%Lincoln Int%’, ‘%Little Lincoln%’, ‘%- Lincoln%’, ‘%(Lincoln)%’, ‘%Lincoln’, ‘%UNL%’, ‘%Middlebury%’, ‘%Nova net%’, ‘%Novanet%’, ‘%Odyssey%’,‘%Odware%’, ‘ODY%’, ‘%(OD%’, ‘%Edmentum%’, and ‘%Plato%’.
• TSDL Local Course Title Field References Common Generic Labels for Online or Virtual Learning. Searching the local course title field for common references to online, distance, or virtual learning yielded 1% (4,859) of the virtual enrollments. The wild card search criteria for common generic labels for online or virtual learning consisted of the following: ‘%Online%’, ‘%On-line%’, ‘%On line%’, ‘%onl’, ‘%onli’, ‘%onlin’, ‘%- OL’, ‘%-OL’, ‘%O/L%’, ‘OL %’, ‘%STW%’, ‘%E-Learn%’, ‘%E-LRN%’, ‘%Virtual%’, ‘%- virt%’, and ‘%- DL’.

Michigan Virtual Students

In past years of this report, data on Michigan Virtual students were presented with two known issues. First, because Michigan Virtual is not a school entity and its data are not part of the state data sets, Michigan Virtual and CEPI used a matching process to attempt to tie the records together. This was always an imperfect matching process where some students taking Michigan Virtual courses were not able to be linked in the data set and, therefore, under-reported on students taking courses with Michigan Virtual. The second issue was that even when the process yielded a match for a student, it could not be determined which virtual courses the student took with Michigan Virtual as opposed to another course provider. Faced with that dilemma, the approach taken was to count all virtually delivered enrollments for such students as being provided by Michigan Virtual, recognizing that not all were actually delivered by Michigan Virtual. Given these issues, reporting on Michigan Virtual students as a subset within the report ended with the 2017-18 report.

However, Michigan Virtual’s Annual Report: 2019-20 does contain information on its student enrollments. For instance, data are provided on the number of students served (over 18,100), districts served (over 60% of LEA Districts), enrollments served (over 32,600) and its pass rate (77.8%). For more information about Michigan Virtual enrollments for the same school year as this report, please see that annual report.

Pass Rate Calculations

For this report, the pass rate was calculated based on the values recorded in the “Completion Status” field. For more information about the Completion Status field, including definitions for each status, see page 350 of the Michigan Student Data System Collection Details Manual Version 4.0. Column one of Table A1 displays the various statuses reported by schools for the virtual enrollments.

Throughout this report, the pass rate calculated simply represents the percentage of virtual enrollments with a status of “Completed/Passed.” Notice that the percentage of enrollments with a “Completed/Passed” status in Table A1 matches the statewide pass rate. This pass rate formula remains consistent with past reports. Please keep in mind that calculating the pass rate in this manner will result in the lowest possible percentage.

To illustrate why this is, consider the completion status of “Audited (No Credit Issued).” These virtual enrollments are not “failures” per se; however, including these enrollments in the total counts adds to the formula’s denominator without impacting to the numerator, the effect of which is to lower the percentage of other completion statuses including “Completed/Passed.” Another example is enrollments with a completion status of “Incomplete.” About 10% of the virtual enrollments in this report were classified as “Incomplete.” As such, they are treated in the report’s pass rate formula as zero passes, even though some may eventually be awarded a passing status. Finally, it is unclear how to best treat enrollments with a “Withdrawn” status. For instance, 5% of the virtual enrollments in 2019-20 were marked as “Withdrawn/Passing,” meaning that the student was passing the course at the time the student was withdrawn. Should these enrollments be counted as failures? What about students whose enrollments were marked as “Withdrawn/Exited” (8% of the virtual enrollments)? There is no way to determine whether that exiting occurred in the first few weeks of class or the final weeks of class. The data do not provide insight into whether the student was re-enrolled in a different course or whether it was too late for re-enrollment in a credit-bearing opportunity for the student.

The research team raises these issues because they represent questions for which there are no definitive answers. In the end, the team decided to report the pass rate as the percentage of all virtual enrollments that were reported as “Completed/Passed.” To provide readers with a better idea of the impact of this approach, additional data tables are provided in Appendix F to allow interested readers in drawing their own conclusions and calculating their own formulas for many of the pass rates reported.

Data Limitations

Because of the methodology described above, some enrollments are counted as virtual in this report that should not be – either because they were mistakenly marked as virtual by the school and/or because the local course title searches implemented by the research team yielded false positives. On the other hand, it is also safe to assume that some enrollments that should have been marked as virtual were not, both because they were not correctly flagged by the school and because the local course title did not give an indication of its virtual nature that aligned with the conventions used in the strategies outlined above. Consequently, the figures in this report should be treated as estimates that, generally speaking, convey the trends observed for the school year.

One final caveat for interpreting the results published in this report: There is clear variability in what schools report to the state as a “course.” Some records align well with reporting conventions outlined by the U.S. Department of Education under their School Codes for the Exchange of Data (SCED) (see the U.S. Department of Education’s School Codes for the Exchange of Data). However, a review of the data suggests that many schools submit course records that may be better described as course units or lessons. Hence, while one district may report a single course for a child, for instance, Algebra I, another school might submit five such records, all with the same subject areas and course identifier codes, but with different local course IDs. Table A2 provides a glimpse into such reporting variability. Consider the part-time schools. For that group, 50% of the students had 11 to 15 courses reported (including both virtual and non-virtual enrollments), but 20% of part-time students had more than that. This issue is pointed out to alert readers that not all courses likely represent a semester- or trimester-length unit.

Appendix – B School Tables

Note: Clicking on the orange hyperlinked table number will return to the section of the report that discusses the table.

Note: Data are not reported (NR) if there were less than 10 schools for that cell or to prevent calculating cell value.

Note: Because some students took courses across multiple entity types, a student may be counted toward more than one type. The total row, however, reflects the number of unique students.

Note: Data are not reported (NR) if there were less than 10 schools for that cell or to prevent calculating cell value.

Note: Data are not reported (NR) if there were less than 10 schools for that cell or to prevent calculating cell value.

Note: Data are not reported (NR) if there were less than 10 schools for that cell.

Note: All Michigan K-12 schools with building codes were used to calculate the state figures.

Appendix – C Course Tables

Note: Clicking on the orange hyperlinked table number will return to the section of the report that discusses the table.

Note: % of Enrolls based on the overall total of 129,014 for this subject area.

Note: % of Enrolls based on the overall total of 117,399 for this subject area.

Note: % of Enrolls based on the overall total of 93,463 for this subject area.

Note: % of Enrolls based on the overall total of 111,056 for this subject area.

Note: An additional 440 enrollments had a course type listed as Advanced Placement, but did not match an AP SCED Code. Similarly, there existed local course titles with AP in the title that did not have an AP SCED Code. Thus, it is very likely the data above underreports the number of students taking AP courses virtually. Pass Rates are not reported (NR) if there were less than 25 for that cell.

Note: Data are not reported (NR) is there were less than 25 virtual enrollments for that cell.

Note: Pass Rate data are not reported (NR) if there were less than 25 virtual enrollments for that cell.

Appendix – D Student Tables

Note: Clicking on the orange hyperlinked table number will return to the section of the report that discusses the table.

Note: Because some students took courses across multiple grade levels, a student may be counted toward more than one grade level. The total row, however, reflects the number of unique students.

Note: A few students had enrollments where their sex was listed as male on some, but female on others.

Note: The total number of students exceeds the 121.900 number because a few students had enrollments across multiple schools where one school listed the student as one race/ethnicity, but the other school reported a different value. The unique total was used to emphasize the true number of virtual students.

Note: The total number of students exceeds the 121,900 number because a few students had enrollments across multiple schools where one school listed the student under a specific poverty status, but the other school left the status unknown. The unique total was used to emphasize the true number of virtual students.

Note: The Virtual Pass Rate – Non-Virtual Pass Rate calculation was run prior to rounding. That rounding effect accounts for what may appear a calculation error.

Note: The total number of students exceeds the 121,900 number because some students had enrollments across categories. The unique total was used to emphasize the true number of virtual students. Also, all Michigan K-12 schools with building codes were used to calculate the state figures.

Note: The total number of students exceeds the 121,900 number because a few students had enrollments across multiple schools where one school listed the student under a specific special education status, but the other school left the status unknown. The unique total was used to emphasize the true number of virtual students.

Note: The total number of students exceeds the 121,900 number because a few students had enrollments in both full-time and part-time virtual schools. The unique total was used to emphasize the true number of virtual students.

Appendix – E State Assessment Tables

State assessments were suspended for the 2019-2020 school year due to the pandemic.

Appendix – F Completion Status Tables

Note: Clicking on the orange hyperlinked table number will return to the section of the report that discusses the table.

Note: UEP = Unique Education Provider

Note: Reportable Programs and Vocational/CTE are not reporting on here because each had less than 10 schools.

Note: Only Race / Ethnicities with 1,000 or more students are reported in the table.