Skip to main content


Welcome, the Hub connects all projects

Voices From The Field


MSPnet Blog: “NEPC’s 2018 report on virtual and blended schools”

See All Blog Posts

posted May 7, 2018 – by Brian Drayton

The NEPC has just issued its  6th annual report on virtual education, which you can get here. This study examines both full-time virtual schools (429 schools, 296,000 students), and full-time blended schools — those with a substantial on-line component as well as “traditional” face-to-face instruction  (296 schools, with 177,000 students).

I think this is a valuable series to follow, because in many states or regions, the push for increased school “choice” has revealed the basic problem that if I decide I want my student to move out of a poorly-performing school, I may actually have no place within practical reach to move her to.  Many choice advocates, in response, suggest that virtual schools are one good answer — and so, like charters, virtual schools are encouraged in many states — one way they are encouraged is by having less oversight than “traditional” schools (See here for a story on the tension around regulation, de-regulation, and school quality in Indiana.  h/t to Hackeducation.com again).

Meanwhile, blended schools are being encouraged in many states as part of the move to personalized learning, as well as to provide access to otherwise unavailable content for students in isolated or low-resourced schools.

In both cases, while private (nonprofit and for-profit) “Educational Management Organizations” (EMOs) are in the minority numerically (a little more than a third of the virtual schools for example), they have higher numbers of students per school (on average), and as a result have almost 2/3 of the students in such schools.  For profit schools tend to have higher class sizes — for example,  for blended schools, “Blended schools enrolled an average of 394 students, but blended schools managed by for-profit EMOs had a far larger average enrollment of 1,288.”

The authors believe that the trend towards poor teacher/student ratios in virtual schools is an area for urgent attention and remediation — three of their policy recommendations are related to this, including one that surprised me, because it suggests some issues worthy of deeper investigation:   Policymakers should

Require that teachers employed by virtual schools, and not parents, take primary responsibility for students’ education. The widely practiced corporate model instead largely relies on the parent as teacher and provides contracted teachers with insufficient time to interact with students and to provide support for those who struggle or drift away.

In the case of both virtual and blended schools, NEPC continues to find relatively little research on the innovations, but what information is out there, they have once again collated and examined.  Moreover, even at the basic level of data collected by the licensing agency, data are hard to come by: “Many states continue to have frozen accountability systems or to have implemented new systems that do not include an overall rating.”

What data there are, are sobering:

Virtual schools continued to underperform academically, including in compar- ison to blended schools, although the margins were much closer this year than last. Overall, 36.4% of full-time virtual schools and 43.1% of blended schools received acceptable performance ratings…. On-time graduation rate data were available for 247 full-time virtual schools and 152 blended schools. The graduation rates of 50.7% in virtual schools and 49.5% in blended schools fell far short of the national average of 83%

It is interesting to note that when districts are responsible, performance measures are more encouraging:

Among virtual schools, district-operated schools performed far better based on school performance ratings (53.8% acceptable) than charter-operated schools (20.7%).

The largest proportion of virtual schools are secondary (26% high school), while an even larger proportion of blended schools are high school (41%), and in both varieties, an uncertain but additional proportion are actually K-12 schools.  Virtual schools tend to be whiter (66% of students are “white non-hispanic” vs. the national mean of nearly 50%) (blended schools have a compostion much more similar to the national averages, both with respect to ethnicity and to special needs) .  The authors’ comment on the demographic characteristics of virtual schools makes clear how complex the story may be, and how much more data are needed to really understand what’s going on:

The fact that minority low-income families may have less access to technology may help explain underrepresentation of these groups, even though many of the virtual schools loan their students computers and often pay for internet access. There are other possible explanations for the overrepresentation of White students in these schools, such as White flight by urban families or the fact that virtual schools often pres- ent the only viable form of school choice in rural areas where minorities are less prevalent. These possible explanations warrant further exploration to determine whether they can ex- plain underrepresentation of some ethnic groups in virtual schools.

As for measures of academic performance, the news continues discouraging for virtual schools, based on the available studies:

The first decade of the new millennium provided little research into full-time virtual and blended school student achievement at the K-12 level, and results of existing research were not positive…To date, the second decade hasn’t produced much research either, and the scant research conducted has produced mixed results, with most findings being either neutral or negative…This report has consistently found virtual schools not performing well.

Although the indicators for blended schools seem better, the authors note that our ignorance about the performance of blended schools is extensive, with the lion’s share of research focusing on virtual schools.  I expect that this may change, with the increasing investment in “personalized learning,” since it seems more and more common for schools to provide at least some of the personalization through the use of on-line course work and other resources.

The authors make several recommendations for policy, and specific areas in which further research is needed.  Their final conclusions, however, lead this reader, at least, to the conviction that in this as in many aspects of the current “reform” measures, we are conducting “experiments,” but not well-designed ones, and proceeding in the faith that this is a good way to go.  The kids, the parents, and the teachers are enlisted (both in the virtual schools, and in the real schools), with varying degrees of informed consent, but too much of the insight being gained is provisional, post hoc, or anecdotal.    The study’s final recommendations:

Given the rapid growth of virtual schools and blended schools, the populations they serve, and the relatively poor performance of virtual schools on widely used accountability mea- sures, four final recommendations are offered.

  • Policymakers should slow or stop the growth in the number of virtual schools and in the size of their enrollments until the factors responsible for their relatively poor performance have been addressed.
  • Policymakers should carefully and continuously monitor the performance of full- time blended schools since the evidence base is still weak.
  • Authorities charged with oversight should specify and enforce sanctions for vir- tual and blended schools that perform inadequately.
  • State agencies should (1) ensure that virtual and blended schools fully report data related to the student population they serve and the teachers they employ, and (2) make every effort to assign all virtual schools an overall school performance rating and explain each missing rating.

Note:  The opinions expressed herein are those of the author, and not necessarily shared by MSPnet, TERC, or the National Science Foundation.

Blog comments have been archived, commenting is no longer available.
This blog post has 1 comment.

Research Designs

posted by: Betsy Stefany on 5/8/2018 1:41 pm

I tend to agree with you on your conclusion concerning the design challenge and this report.
Creating a science type design has always felt problematic from my perspective when expecting individual change. As I deal with projects that are questioning the engagement of individuals and how they embrace STEM, change is more than an intervention. The goal is to accept and engage with the introduction of an innovation. The terms and processes by reflections may not serve to transfer the actual individual experience. Also the large group methods seem to be developed to find a mathematical level or average. Designing to engage in truly personalized learning would seem to me to mean an agreement by researchers on the value of ensuring they are measured by personalized evaluation. The virtual and blended schools are approaching the full system approach but as you observe from the hard to categorize expectations of individuals, students as well as their parents.

Hope you will stop by our STEM-for-All video and watch us share several approaches to technology integration and ask our evaluation researcher about her approach to the multiple projects that evolve from grassroots up development of personalization through data.