Nearly every study of virtual school performance has found their performance to be lacking. The most recent and definitive study was conducted by the Center for Research on Education Outcomes (CREDO) at Stanford University in 2015. CREDO controlled for the unique characteristics of students enrolled in virtual charter schools by comparing their performance to a “virtual twin,” a student with the same demographic characteristics and similar prior achievement enrolled in a traditional public school. CREDO declared,

Across all tested students in online charters, the typical academic gains for math are -0.25 standard deviations (equivalent to 180 fewer days of learning) and -0.10 (equivalent to 72 fewer days) for reading.

Considering that a typical school year consists of 180 days of instruction, many have observed that it’s almost as if these students had no schooling whatsoever. As a colleague of mine remarked, “They could have learned just as much playing video games for a year.”

Now, not all virtual schools are charter schools, but many are. According to the National Education Policy Center, virtual charter schools make up approximately half (52 percent) of all virtual schools in the United States, and together they accounted for 84 percent of full-time virtual school enrollment.

A Growing Body of Evidence

Virtual school operators respond to the CREDO data by asserting that their students are different and that their troubles in traditional schools invalidate these comparisons—a point of view reiterated by my forum partner, Tom Vander Ark. Yet CREDO looked at these students’ performance in other settings, and found that the same students showed stronger performance both before and after their tenure in virtual schools.

But if we’re having a debate about an individual study, we’re having the wrong debate. The conclusion that virtual schools aren’t working is based on an accumulation of evidence, all of which casts doubt on virtual school performance.

Over the years, a number of studies (from Minnesota, Colorado, Pennsylvania, Ohio, and Texas) have documented dismal outcomes in virtual schools, including low course-completion rates and higher-than-average school dropout rates. While many of these studies are descriptive and less definitive than the CREDO study, they all point in the same direction.

It’s important to note that these studies—and my commentary—refer to schools that deliver 100 percent of their curriculum online. There are hundreds if not thousands of other schools that are providing a blended learning curriculum or online supplemental courses. Those schools differ from schools that are fully online and may offer more hope for combining the best of both digital and in-person learning.

So Many Students So Poorly Served

Evergreen Education Group, a leading consultancy in the digital learning field, estimates that 275,000 students are enrolled full-time in virtual charter schools. This is a huge number of students. Not only are most receiving an inadequate education, but also the sheer size of the group skews the perception of performance of charter schools as a whole.

Those 275,000 students represent more than the total number of charter school students enrolled in the nation’s 10 largest charter school communities, according to the National Alliance for Public Charter Schools. Those 275,000 students exceed the 271,000 students that Bellwether Education Partners estimates are enrolled in all of the high-performing charter management organizations in the United States.

Take a moment to contemplate that fact: The positive impact of years of work done by thousands of educators to build networks like KIPP, YES Prep, Achievement First, Noble, Mastery, Uncommon, Aspire, IDEA, Harmony, and others is literally negated by the performance of virtual charter schools.

There are solutions at hand to help local leaders and policymakers who want to take corrective action. All we need now is the will to implement them.

It’s Time to Act

In its study, CREDO found that “two percent of the online charter schools outperform their comparison schools, 32 percent perform no differently, and 67 percent have weaker growth than their comparison schools.” Those data suggest that authorizers should take immediate steps to close many of the virtual charter schools in existence today. The closure of schools that persistently fail students is one hallmark of the charter school philosophy—and authorizers need to enforce it.

Despite the dismal performance of virtual charter schools to date, there’s a need for virtual schooling of some type in the future. As the Evergreen Education Group notes, thousands of students “are attending these schools because they have medical or behavioral issues, are engaged in a time-consuming pursuit such as arts or sports, or have not been academically successful in a physical school and are seeking a different mode of instruction.”

I am not calling to eliminate virtual schooling, but to create an environment in which it’s successful for students and taxpayers. This begins with the recognition that aspects of the current model aren’t working:

• Virtual schooling is not the right fit for all students. Only some students have the parental support or self-discipline needed to be successful in these programs. Unlike charter schools, which must be open to all, enrollment criteria should be established for virtual schools based on factors proven necessary for student success. This could include ensuring students and their parents are made fully aware of the demands of the virtual school model and identifying considerable adult support.

• Funding. Charter schools typically receive a full year of funding for each student they enroll. Yet many virtual school students stay enrolled for only a few months, by design. This is a waste of taxpayer dollars. Virtual schools should be funded, not on how many students they enroll, but on how many students successfully complete their coursework. Partial payment for partial completion can address the funding issue for students who attend virtual schools only briefly.

• Accountability. Charter schools are accountable for their performance on state standardized tests administered each spring. Yet virtual school students come and go at all times of the year, leaving the public with no data on their performance. Fortunately, virtual schools have an enormous amount of detailed data on the progress of each and every student. Right now, they are not making that data public. They should—and they should be held accountable for those results. Alternatively, there are many reliable interim assessments to measure student growth, such as the Northwest Evaluation Association’s Measures of Academic Progress (MAP) that can be given every nine weeks and that measure growth no matter how long a student attends a virtual school.

The tough policy changes needed to address these issues may prove incompatible with state charter school laws in the United States. Should that be the case, legislatures should pass new legislation specifically designed for the unique attributes of virtual schools.

We Know What to Do

To provide students with better options in the future, authorizers need to close virtual charter schools that are persistently failing. And state legislatures need to pass new legislation that establishes unique ground rules for virtual schools.

They need to do so immediately. The education of 275,000 students depends on it.

Greg Richmond is the President and CEO of the National Association of Charter School Authorizers.

This is part of a forum on virtual charter schools. For an alternate take, see “Online Charters Expand Learning Options,” by Tom Vander Ark.