Study of online passing rates in Michigan suggests that online learning is not for everyone

A previous post looked at general findings of the recently released virtual learning study in Michigan. In this post I review more deeply one key aspect of the study: the rates at which students are passing online courses, and implications for how online courses and schools are made available to students.

The Michigan report notes that the “completed/passed” rate for virtual course enrollments in school year 2013-14 dropped three percentage points from the previous year to 57%. This is significantly lower than the 71% rate at which virtual students passed traditional (non-online) courses. The researchers believe that some of the non-completed courses at the time of data collection would eventually be completed and passed, raising the overall completed/passed percentage to perhaps around 61%. Still, the pass rate for virtual courses is undoubtedly lower than pass rates for all students in traditional courses, and lower than pass rates for virtual students in non-virtual courses.

Digging deeper into the data, as the Michigan researchers did, reveals other important distinctions:

“This average [of passing rates], however, may not be of great value in understanding what is happening with virtual learning…over half of the students who took one or two virtual courses passed all of them. Over 40% passed all of their virtual courses if they took three or four. While the percentage of students who passed all of their virtual courses trended down to around 30% as the number of virtual enrollments they took increased, the percentage of students who passed none of their virtual courses tended to be fairly consistent; whether students took one, two, three, four, or even five virtual courses, about 30% failed all of them…Virtual Learners tended to be highly successful or highly unsuccessful; few students fell in between. This pattern of high or low success was similar for school measures. Many schools had high success rates for their virtual learners; too many did not. (Emphasis added.)

Very few educators—even those who are advocates for online learning—believe that online courses and schools are a good fit for every student. Colleges and universities know this as well. A Google search for the term “is online learning right for me?” returns page after page of colleges and universities that have posted information or short quizzes meant to be used by students to determine if online learning is the best option for them. Post-secondary institutions want students to self select into—or out of—online courses and degree programs.

K-12 education presents a different and challenging set of issues related to this question. By law (and ethically as well), public K-12 educational options must be available to all students in almost all cases. Public online schools and online course providers in course choice programs cannot turn away students except in cases where they have reached capacity, and then they must use a lottery system to determine which students they accept. The question is whether they can counsel students to help students and families make the best decisions for all students—whether these include an online component or not.

This is not an easy question, as it is mixed with contentious ethical and legal issues. On one hand, the idea that schools would help students make the best decision makes sense—who better to know the aspects of online courses and schools than the schools themselves? On the other hand, any attempt to allow schools or course providers to counsel students out of taking online options runs the risk of having providers attempt to attract only the best students, and not serve students who have characteristics associated with lower rates of student achievement.

It seems to me that these issues are rarely discussed publicly, perhaps because they can be so contentious. But practitioners know, and the Michigan study affirms, that online courses and schools are not for everyone. Students, families, educators, and policymakers would all benefit from an honest exploration of how to ensure that students are counseled into the best options for them, while protecting against potential abuses of such counseling.

Michigan study provides detailed online learning data; shows student attributes and growth in online enrollments

The Michigan Virtual Learning Research Institute, which is a center at the Michigan Virtual University®, has recently released its latest Virtual Learning Effectiveness Report. The study reviews virtual learning in Michigan during school year 2013-14.

The report provides summary information showing the number of virtual courses taken by students in Michigan:

  • 76,122 K-12 students in Michigan took one or more virtual courses during school year 2013-14, an increase of about 38% over the prior year. Some of the growth appears to be a result of data collection changes for this report, but the researchers believe that most of the increase represents real growth.
  • The total number of online course enrollments was 319,630, an increase of about 73% over the prior year.
  • Most enrollments (68%) were from local school districts. Online charter schools accounted for 27%, and Michigan Virtual School accounted for 5%.

Another set of data points is quite a bit more complicated and compares course pass rates for students in online courses compared to students in traditional courses. According to Dr. Joe Freidhoff, Executive Director of MVLRI and author of the report:

“What we found was that students who did not take any virtual courses in 2013-14 passed their courses (let’s call them traditional courses) 89% of the time…When virtual learners took traditional courses, they only passed those courses 71% of the time. This was an important finding because if virtual learners as a group were the same as non-virtual learners, then we should have found that number to be the same. Instead, we found a sharp drop off indicating that the students taking virtual courses at present, tend to perform below average. When virtual learners took virtual courses, their pass rate dropped to 57%, though in reality this percentage will likely end up higher as 8% of the virtual enrollments were marked as “Incomplete” at the time of the study and many, if not most of them, have likely been reclassified to another status. This issue is mainly due to the prevalence of virtual learning over the summer and the fact that we needed to collect the data for the report prior to some schools reporting final data for the summer.

Despite the impact of “Incompletes” in the data, it certainly is not enough to change the conclusion that virtual learners themselves did worse in their virtual courses than they did in their traditional courses. For opponents of virtual learning, this likely will be used as evidence that virtual learning does not work. This question, “is virtual learning better or worse than traditional brick and mortar learning,” is a red herring. The data in this report, as has been found in plenty of other research, shows that virtual learning can be successful, but not always so. Speaking as a researcher, what we seek to understand is the differences between those that were successful and those that were not so that we can improve learning outcomes for more students.”

Summarizing this finding, Dr. Freidhoff states

“The data seem pretty clear that, at present, virtual learning is being used by schools for students who are struggling in their traditional setting rather than to enable average and above average performing students to go farther faster with their education.”

MVLRI has also recently looked at student demographics in online schools, comparing students in online schools in Michigan to national averages. According to the NEPC report that Keeping Pace reviewed recently, compared to national averages “virtual schools had substantially fewer minority students, fewer low-income students, fewer students with disabilities, and fewer students classified as English language learners.” Online schools in Michigan appear similar in some ways and different in some ways from the national averages:

“The national trends that more females are being served by virtual schools, fewer minority students are enrolled in virtual schools, and virtual school students are less likely to be classified as English Language Learners appear to be true for virtual schools in Michigan for the 2013-2014 school year. In contrast to the national trend, Michigan appears to have virtual schools serving higher percentages of low-income students than the statewide average and four of the seven virtual schools have higher rates of students with disabilities compared to the state average.”

Although the report discusses some concerns with the data, the study demonstrates the value of state reporting of online course enrollments. Very few other states are coding enrollment data to show online courses, and there is no other state for which a report as comprehensive as this one exists.

A review of the latest report on virtual schools from the NEPC

The National Educational Policy Center (NEPC) reports on virtual schools and related issues fairly regularly, and has recently issued its 91-page report Virtual Schools in the U.S. 2015: Politics, Performance, Policy, and Research Evidence.

Similar to past NEPC reports (see for example comments here and here), I find that the latest study has a mix of useful information, recommendations with which I agree, and recommendations with which I disagree. In some cases statements and recommendations are unsupported by the evidence that the report presents.

Among the findings and recommendations that I appreciate are the following:

  • “[S]tate education agencies and the federal National Center for Education Statistics clearly identify full-time virtual-schools in their datasets, distinguishing them from other instructional models.” This would be an incredibly valuable. In the last couple of years an analogous effort has taken place in post-secondary education, and a similar approach would benefit research and policy in K-12 education.
  • “Develop new accountability structures for virtual schools” and “promote efforts to design new outcome measures appropriate to the unique characteristics of full-time virtual schools.” Keeping Pace research supports the concept of new accountability structures for online schools. We have explored this issue in School Accountability in the Digital Age and in two separate blog posts.
  • “Develop a comprehensive system of summative and formative assessments of student achievement, shifting assessment from a focus on time- and place-related requirements to a focus on student mastery of curricular objectives.” Implementing this recommendation would be challenging on many levels, but worthwhile. In our as yet unpublished research into examples of success in blended learning, we are finding schools that are using NWEA MAP or similar sets of assessments in order to move closer to this objective.
  • “Define new certification training and relevant teacher licensure requirements and continually improve online teaching models through comprehensive professional development.” Although I’m not convinced that new certifications are ideal, there is no doubt that extensive professional development is a key component of successful programs, such as in Clark County, NV. Keeping Pace research also suggests the need for changes to teacher licensure, although the Keeping Pace and NEPC recommendations aren’t exactly aligned.

Areas where I believe the evidence doesn’t support the NEPC findings include:

  • In the introduction, the report characterizes the “virtual school expansion” as “fast-growing.” The same introduction includes this statement: “A total of 400 full-time virtual schools enrolling an estimated 263,705 students were identified, an enrollment increase of some 2,000 students since last year’s report.” NEPC’s numbers suggest a growth rate of less than 1%–which is hard to reconcile with the statement that the sector is “fast-growing.” The numbers that Keeping Pace 2014 reported are larger (about 315,000 students and a 6.2% growth rate), but they are still very small relative to K-12 education as a whole.
  • The report discusses, fairly extensively, the number of introduced legislative bills related to online schools that failed, and ascribes importance to these failed bills. In some new areas, such as data privacy, the number of introduced bills can be indicative of interest and predictive of future activity. For a topic such as online schools that is fairly mature, however, proposed or introduced bills are not necessarily predictive. At the very least, bills that pass are orders of magnitude more important than those that fail, and the long discussion confuses these different categories.
  • Finally, in calling for new research and accountability systems, the report suggests that current methods of measurement are not ideal. Yet it appears that the authors base some of their findings and recommendations on the very metrics (e.g. measures of student proficiency) that they deem inadequate.

The report is 91 pages long, and includes many additional findings and recommendations. One area that I’ll review in an upcoming post will be the exploration into numbers of minority, low-income, and special education students in online schools—which is among the most valuable areas of the report.

Digital Learning Snapshot: Horry County Schools, Conway, SC

This is the fourth in a series of blog posts about digital learning innovation in school districts. See our first post, introducing the series, and posts on Clark County, Nevada and Washington DC Public Schools.

Horry County Schools (HCS) offers and is further developing a set of digital learning options that include the following:

  • Continuing roll-out of district-wide Personalized Digital Learning, extending from middle school in early 2014 to high school in fall 2014, and expanding to grade 5 in all 27 elementary schools in 2015 with plans to add grades 3-4 in the future.
  • A low-performing, high-poverty middle school (Whittemore Park) that the district is turning around using blended learning in a competency-based learning setting.
  • The Horry County Virtual School (HCVS), which provides supplemental online high school courses to district students. It offers both original credit and credit recovery courses and totaled 3,500 course enrollments in SY 2013–14.

Several efforts in place in the district have supported the move to digital learning. HCS has traditionally relied heavily on data to differentiate classroom instruction, using the Northwest Evaluation Association’s (NWEA) Measures of Academic Progress (MAP) as well as state assessments to track student progress. Regular administrative meetings were often used as professional learning opportunities to discuss data and student progress, and teachers and administrators frequently collaborated to make meaningful use of student data to improve outcomes. During these collaborations, district staff decided that blended learning was the logical “next step in the journey.”

In 2013, HCS decided to move to personalized digital learning for all of its students over the course of the next three years. Because district leaders believed that teachers and students should have a role in designing how digital learning would be implemented in each school, design teams of students, teachers, and administrators were formed at schools across the district, and implementation varies between schools.

To augment the district’s efforts, it received a Next Generation Learning Challenges (NGLC) grant in October 2012 to help its efforts to use digital learning to turn around Whittemore Park Middle School. In fall 2013 the school implemented the iCAN (individualized, college and career readiness, aspirations of students, and network of support) model designed to blend core subjects and provide an increased level of student support. Fluid groups of about 100 students meet with four academic teachers for 300 minutes each day. About 75% of that time is digital learning instruction. Each week students also meet in advisory groups that are static and experience a variety of exploratory classes. The school is located in an urban setting where over 85% of the students participate in the free and reduced lunch program.

Digital learning content and tools

A wide array of digital content is used in in the district’s virtual school program (Horry County Virtual School), and in its personal digital learning based schools. All core middle school subjects use digital content. The district started with English language arts and math and has expanded to social studies and science, where digital textbooks are often utilized. The virtual school uses a mix of vendor-provided courses with those developed in the district. Most of this content is integrated with the district’s learning management system. Students in grades 5–12 are being issued tablets, though the devices are different for high schools. The high school device has an integrated keyboard, while the elementary and middle school device does not. The school board has approved up to $600 per student per device.

Teaching and staffing

Teachers and administrators are participating in professional development to understand how to teach effectively in a blended learning classroom. In addition, six digital integration specialists were hired to support existing staff with the blended learning implementation. Each building has a curriculum coach who collaborates with digital integration specialists and content learning specialists to support blended learning in the classroom.

Most teaching at HCVS is through 15 part-time teachers, and for some low enrollment courses the teaching and content are supplied by an outside provider. The district’s executive director of online learning and instructional technology leads HCVS. A learning specialist for online learning and curriculum monitoring supports the director and handles day-to-day operations and enrollment requests. An administrative support/clerk position also plays an integral role in supporting HCVS administrators, teachers and students. These are the only three positions totally dedicated to HCVS. HCVS also receives as-needed support from the district’s curriculum staff and the technology staff; however, these positions are not dedicated to HCVS

Responsibility for the district’s Personalized Digital Learning Initiative spans several departments. The chief academic officer and the executive director of online learning and instructional technology serve as project leads on the instructional side. The chief accountability/technology officer and executive director for technology serve as project leads on the hardware side. Although these positions already existed at HCS, the board approved six new positions to assist teachers in making the shift to blended learning. The six new digital integration specialists report directly to the executive director for online learning and instructional technology.


General fund monies are used to cover most of the costs of implementing and executing blended learning. Student devices and building infrastructure are purchased through a capital fund supported by a $.01 local sales tax.


HCS is putting into place an impressive district-wide plan for personalized digital learning that builds on previous use of data by district schools, widespread adoption of tablets for students in grades 3–12, a whole school turnaround based on digital learning, and a district virtual school. Digital options are focused on the core subject areas at each level. The district relies on school leadership teams from each level to assist in selecting and adopting digital content options across schools.


New report from iNACOL explores Performance-Based Funding

iNACOL has just released Performance-based Funding & Online Learning: Maximizing Resources for Student Success. Evergreen, along with Augenblick, Palaich and Associates, Inc. (APA), collaborated on the research. As part of the research we convened many state education policy leaders and school finance experts, who provided exceptionally valuable insights into school funding.

From the iNACOL press release:

“This study examines the variety of funding models for K-12 online learning, explores adequacy costs, and outlines the guiding principles of performance-based funding to reward outcomes. As state policy makers consider performance-based funding options in education, this report provides recommendations and literature to expand the research base in support of equity, closing the achievement gap and increased student outcomes.

Susan Patrick, President and CEO of iNACOL, said, ‘This report asks the question, What if education funding was not based on seat-time, but on rewarding outcomes in student performance? It highlights differences in how online learning programs are funded in states and explores policy principles to ensure equity and adequate funding levels. We look forward to working with state policy makers who wish to explore performance-based funding further.’… Because the ability for states to implement performance-based funding varies, the guidance offered in this report provides flexible steps that they can take, including variable starting points to allow state leaders to adapt funding policies to local contexts.”

In addition to the policy recommendations surrounding performance-based funding, the report also contributes valuable supporting information, including the following:

  • A discussion of how funding for online schools compares to state education funding averages (beginning on page 16)
  • An exploration of funding for supplemental online programs, and in particular state virtual schools (beginning on page 19)
  • The most recent adequacy cost study, which was conducted by APA as part of the research for the report (beginning on page 26).

The cost study found that “Using personnel salaries in Colorado and Pennsylvania under this costing-out approach, the cost of the full-time online school (that is resourced to bring all students to college- and career-ready success) is between 93% and 98% of a traditional school cost.” It also found that “the cost of providing supplemental online courses would be about $600 per student, per semester course.”

For anyone interested in hearing more, iNACOL will be hosting (and I will be joining) a webinar to discuss the research findings on April 8 at 2ET. To register see

WordPress SEO fine-tune by Meta SEO Pack from Poradnik Webmastera