Blended learning can save some money; online learning can save a lot. That’s the conclusion of a working paper—The Cost of Online Learning—from Fordham’s Creating Sound Policy for Digital Learning series. A talented team from Parthenon lead by Tammy Battaglino wrote the report. They estimate “that full-time virtual schools cost, on average, $6,400 per pupil, compared with $8,900 for blended schools and $10,000 for traditional brick-and-mortar public schools.”
Online learning providers won’t agree with the projected 35% savings. They would say that with staffing ratios not all that different from traditional schools there is some opportunity for cost savings—maybe half of what Parthenon came up with.
Real costs also depend a lot on the clients. It may cost twice as much to serve over-aged and under-credited youth with multiple risk factors than it does to serve motivated well-supported students.
The whole subject of what stuff costs is foreign to education—districts spend what they get whether its $6000 or $26,000 per pupil. There are probably three ways to approach the subject of cost:
- compare: start with average expenditures and add/subtract from there
- zero-based: build up costs by item justifying each component
- return on investment: compare outcomes per dollar invested
The Parthenon team used the first method and outlined what blended and virtual models would do without. A zero-based approach is more work but can be a useful budgeting or design exercise.
I was on a panel at a Colorado Succeeds event this morning with Marguerite Roza, the top expert in school finance, and she recommended the ROI approach, best illustrated in a Center for American Progress report, that compares district outcomes to expenditures. The report nods in this direction by suggesting that “new technology‐rich education models will need to be evaluated based on their productivity.” As education adopts ‘Big Data’ we will be able to measure and compare the results not just at the school level but for each unit of study.
Cost Components. The study looked a couple components in five cost categories: labor, content, technology, operations, and student supports.
On technology, the $1200 per pupil cost for virtual school technology may be accurate for two years ago but it will be lower this year. The $500 per student quoted for blended models seems low—it may be accurate for current pilots but as blends mature they will spend as much on technology as virtual schools. In fact, as they move from 3:1 to 1:3 (kids: devices) blends will probably cost more than virtual schools even with BYOD policies.
The report noted that costs differ depending on whether schools purchase turnkey, purchase components, or build curriculum and tools. What the report didn’t say is that old content is flat and sequential and new stuff doesn’t work together very well yet. There should be a couple of robust learning platform ecosystems two years from now.
Like me, the report is optimistic that technology will take some of the drudgery out of teaching (grading, reporting, etc) and will allow teachers to spend more time with kids on higher order skills.
Part time? The paper should have considered the costs of providing part time and course-by-course virtual enrollment for states planning to follow the access recommendations of Digital Learning Now. In the last legislative session Utah expanded part time access but cut the reimbursement to $750 per course—a level at which it is difficult to staff adequately.
The cost of offering individual courses is less than one sixth of a full time enrollment. Individual course offerings don’t require transcript management, guidance, and full student services. Let’s say those services cost 15% of the estimated $6400, that would imply individual course costs of about $900.
Conversion. Transforming traditional schools to blended models will be difficult politically, technically, and financially. The report notes that, “Almost all new virtual and blended schools to date have been start‐ups; very few schools have attempted to convert from a traditional school to either a virtual or a blended model”
Districts have little ability to aggregate capital for productivity seeking capital expenditures. Some districts can pass tech levies. A few foundations will support conversion work. But in many cases, districts will need to bootstrap their conversion to blended models by phasing work over several years and leasing laptops. Fortunately, access devices are getting cheaper. Blended staffing models can be structured to capture savings.
Funding. While this paper only dealt with cost but it’s worth mentioning that full and part time online learning demands a new school finance system—the right level of funding with the right incentives and conditions.
In the last legislative session, Washington State passed a misguided bill that cut reimbursement rates for virtual schools by 15%. The Federal Way Internet Academy (a school launched when I was superintendent there) received about $5000 per student last year and expects about $4000 per student this year—a ridiculous level of funding if you have expectations of quality. IA will try to make ends met by selling summer school and out of state tuition.
Rather than a dumb across-the-board cut, the state should start practicing smart procurement. For example, in partnership with districts, a state could issue an RFP for AP, STEM, and foreign language courses. Districts could expand course offerings with consistent quality and probably save 50% for these expensive specialty courses.
Smart procurement and performance contracting are future of school finance and quality assurance. Digital Learning Now Element 9 summarizes that funding should be weighted, portable, and performance-based. That means higher risk factors bring more funding, money follows students to the best option, and a portion of funding should be paid out when students complete a grade/course with successful level of achievement.
The Parthenon team suggests that funding systems that provide some incentive for productivity would prod schools and districts to “seek out and adopt higher‐productivity solutions.” The authors, “expect to see the development of school models that both redeploy human capital and technology in ways yet to be envisioned as well as create efficiencies for the system.” That will only be true at scale if funding systems provide incentives for attendance, completion, and achievement.
As noted above, the real cost of producing college and career ready graduates depends on the student population and the risk factors they bring to school. As Dr. Roza said this morning, we need to align state funding systems with our academic goals and create incentives for the services we want to see.
This blog first appeared on Huffington Post