Blending, Upending

Is blended learning the disruptive innovation of K 12 reform?

Today’s most promising experiments in K–12 education combine computerized instruction with immediate assessment and feedback, while carefully linking students to the best traditional classroom teaching practices. It’s called “blended learning”—with the mix of digital and human elements constituting the “blend.”

The big new push in these schools is to make the most of teacher time by deploying educators as tutors and mentors who focus on the precise concepts that each student is missing, without holding back those who are ready to move ahead, and without abandoning those who need supplemental instruction on concepts they haven’t yet mastered. Handling the grunt work of simple instruction and then uncovering these competencies and gaps through online exercises is how the computers help the teachers.

This dramatically different style of instruction attempts to optimize the combination of empowering technology and human touch. And there are hints that it might be not only a godsend for students who often get lost in today’s mass-lecture model of teaching, but also an answer to the spiraling costs of conventional schooling, which have become such a drag on our families and communities.

There are several different models for organizing blended learning. In one, all students rotate in groups among different learning stations: computer work, then teacher-led instruction, then collaborative activities. KIPP Empower Academy in Los Angeles is an example of a school that has had good results with this method. In another, individual students receive the bulk of their instruction online, with roving aides available for quick support, and periodically rotate into follow-up sessions with master teachers working one-on-one or with small groups. Carpe Diem Schools in Yuma, Arizona, is a pioneer of this structure.

Other students mix traditional and computerized instruction either by supplementing a conventional school program with one or more online courses (as is beginning to happen in many public districts), or by enrolling in a full-time online school (as about 250,000 children were in 2011, for instance with the Florida Virtual School) and supplementing that with periodic face-to-face tutoring, enrichment, and examination.

These innovations are very new, and the educators pioneering this more personalized style of teaching are constantly adapting as they discover improved ways of proceeding. But the technologies and managerial insights are improving rapidly, and there are indications they might sharply enhance both student performance and the efficiency of schools. Many philanthropists are excited about this field, and say they’d love to see and fund additional blended-learning institutions.

There are four major ways in which creative educators and donors believe blended learning can have transformative effects: through individualization, improved feedback, teacher effectiveness and satisfaction, and cost control. Models that successfully capture several or all of those advantages could produce dramatic improvements over conventional school methods.


The technology at the heart of blended learning offers unprecedented opportunities for personalized instruction, at one’s own pace and style, in the way that an individual tutor might direct a child. This has all sorts of implications for students—particularly those on the far ends of the performance spectrum.


A view from a Khan Academy video teaching algebra II

With good computerized lessons, students who are struggling to master concepts can be offered almost unlimited opportunities to repeat and reinterpret new material, all without standing out or feeling embarrassed about sidetracking the rest of the class. Likewise, students who have command of a particular subject are able to move onto something fresh or additional without delay or unnecessary repetition. Teachers often feel guilty about wasted potential and bored students. Indeed, that word “guilt” comes up frequently in conversations with educators. But they only have so much time.

Blended learning promises to start solving this problem without segregating students into their own schools or classes by achievement level. As educators at blended-learning pioneer Summit Public Schools in the San Francisco Bay area have discovered, making mixed-ability classes work in practice isn’t a walk in the park, but computerized individualization makes it more feasible than in a conventional mass-lecture classroom. Adaptive-learning software of the sort that Knewton, DreamBox, and many other ed-tech startups are beginning to roll out should be able to meet a child right where he is.

Just because you’re physically sitting in fifth-grade math class doesn’t mean you need to be doing fifth-grade math. If you need to work on second-grade skills, that’s fine, and if you can zoom ahead to trigonometry, that’s great too. As you move forward showing mastery, teachers can keep challenging you. Within a traditional classroom, differentiating like this is extremely hard, even for the best of teachers.

Wendy Chaves, who teaches at Alliance Technology and Math-science High School in Los Angeles, says she was “like two different people”—the Wendy who taught using traditional tools, and the Wendy who intelligently exploited technology. “I was a pretty good teacher,” she says. “Then I came here and realized how much I was doing a disservice to my students. I thought I was good and I realized I wasn’t. I was only reaching 35 percent of my class. That’s not the way I should have been teaching.”

It was a “really harsh realization,” she says. “There definitely needs to be a lot of training with blended learning. I was not prepared for the model whatsoever.” Yet when she started teaching with the benefit of data and computerized instruction, she found that her classes were collectively doing much better, “even though all my students are learning at different rates. They’re not all going to be at the same spot. That’s a huge misconception—that everybody needs to be exactly at the same spot.”

Blended-learning teachers must make their peace with this heterogeneity—which has always existed, but often been hidden in mass-taught classes. With blended learning, students only move on when they’ve demonstrated mastery. “They’re learning at their own speed,” notes Chaves. “You kind of have to relinquish control.”

This individualization is what “we’re most interested in—really personalizing the education for the student,” says Shantanu Sinha, president of blended-learning trailblazer Khan Academy. Much of the media attention originally drawn by founder Sal Khan focused on his instructional videos and the concept of “flipping” the classroom: the idea that students would watch video lectures at home to master the basics, and then work with teachers in class to polish concepts. In Sinha’s latest conception, teachers become “mentors for each student individually. . . . It’s more about personalizing education.” Customizing lessons to each learner, as the increasingly adaptive Khan problem sets make possible, is “the core of why we’re working with schools. . . . We didn’t go in just to use our videos instead of teachers lecturing. There’s nothing truly innovative about that.”

To capture the potential of blended learning, teachers will need to rethink their approach. Sinha notes that some teachers first use Khan Academy materials simply to “augment exactly what a student was going to do anyway.” The teacher still presents the same concepts to all students in the same calendar week. “The teacher is still setting the pace for the student.” When that happens, we’re “seeing improvements, and it’s still helpful—it’s better than the model of worksheets. But it’s not as great as it can be.” The real magic happens when teachers decide that “I will let my students run with it,” says Sinha. “I’m not going to hold you back.”

All children deserve to experience the joy of working hard and then discovering they understand something new. Perhaps the biggest way schools fail bright children today is by letting them think education should be easy. When they finally do encounter challenging work, perhaps at college or in the workforce, they become risk-averse and don’t know how to proceed. They’ve not learned to toil harder and try different approaches until they finally master something that seemed outside their reach.

Just as emphatically, many education reformers are excited about blended learning for its potential to raise the test scores of kids who have fallen behind their peers, and are struggling with average-level material. As Khan says, “I think this problem—the one-pace-fits-all lecture or curriculum—is even more damaging for remedial classes. People’s gaps are all over the place.” Remedial students need to have the holes in their knowledge carefully probed and then backfilled, something computerized instruction excels at, particularly when matched with in-person tutoring informed by up-to-the-minute data reports.

And so, paradoxically, blended learning offers opportunities to eliminate one of the oldest tensions in teaching: Should the instructor teach to the lowest common denominator, address the average level, or reach for peak performance? With blended learning, there is the potential to serve all students well at the same time.

Improved Feedback

Schools are supposed to help children get better at certain skills or areas of knowledge. But how, exactly, do people improve at things?

The old saying is that practice makes perfect, but this can’t be just any kind of practice. Schools that mass assign worksheets every night often fail to see measurable results from such labors other than unhappy kids and parents. Simple repetition isn’t enough. Author Geoff Colvin notes that “extensive research in a wide range of fields shows that many people not only fail to become outstandingly good at what they do, no matter how many years they spend doing it, they frequently don’t even get any better than they were when they started.”

Real improvement requires something called “deliberate practice.” This form of intense training has long been used by virtuoso musicians and athletes to improve, but is remarkably absent from how most of us tackle new skills and ideas at school. It involves figuring out exactly what you know, and what you don’t.

Perhaps your left-handed arpeggios are weak, or you choke up on short putts. You practice these skills over and over again with someone or something right there giving you constant and, ideally, instant feedback. Professional athletes have coaches on top of them if any attempt was better or worse than the last one. They spend many hours watching recordings of their performances so they can learn from each attempt. Skiers work in wind tunnels to hone their form and get instant feedback on what boosts speed. A stand-up comedian gets feedback trying out his material in small clubs before doing a major show. “Deliberate practice is hard,” Colvin writes. “It hurts. But it works. More of it equals better performance. Tons of it equals great performance.”

Needless to say, this kind of deliberate practice is rare in most schools, at least on the academic side. Students do problem sets, but they only see days later if the approach they used was right or wrong. And if it was wrong, they have to wait for time with the teacher to grasp why.


Sal Khan with teacher Courtney Cadwell at Egan Elementary

Feedback requires a lot of work from teachers. Picture a middle-school instructor assigning a grammar worksheet to the 96 students who come through her four English classes each day. If she puts a mere two minutes into each worksheet, that’s more than three hours (192 minutes) of work for her. Unless the school has put money in the budget for grading help, she simply can’t assign more than a few such assessments per week. And the students probably won’t all get feedback from her the next day, let alone instantaneously, on where their grammar is right or wrong. Alex Hernandez, head of the Charter School Growth Fund, points out that “a teacher could stay up 24 hours a day grading papers and not give the feedback on their skills that kids get in ten minutes playing a video game.” In a conventional classroom, the feedback loop is sluggish, if not broken.

The “learning labs” at Rocketship Education, a string of blended-learning academies that started in Silicon Valley and are now expanding across the country, have much in common with those video games. At Rocketship, students are getting up to two hours per day of what looks pretty similar to deliberate practice: work that is right at the student’s level, software that points out exactly what the child is getting wrong, and then the opportunity to practice those skills again until she understands and shows mastery. When musicians get two hours per day of deliberate practice, they start improving dramatically. Is it any wonder that Rocketship children also improve with two hours per day of deliberate practice?

Any academic outcome that involves skills can benefit from practice of this sort. Math is an obvious application, but writing involves grammar skills—and these can be practiced too; new software called NoRedInk, launched from the Imagine K–12 business incubator, applies instant feedback to the matter of subject-verb agreement instead of alien-shooting skills. Foreign languages can benefit enormously from computerized practice in speaking and writing. Basic science skills, which often involve math, or lab procedures can be practiced in the same way that basketball players do sprinting and jumping drills. Philanthropists like the Michael & Susan Dell Foundation are now funding software platforms that integrate data from many different programs so educators can see how their students are doing in deliberate practice sessions, and teach to their weak points.

To be sure, not all of this involves higher-order thinking. But the promise of blended learning is that by having computers take over basic skills practice, doing it quickly and efficiently so students can reap the benefits of a higher volume of practice, teacher time can be preserved for nurturing critical thinking. A teacher who doesn’t have to assign and grade grammar worksheets can spend her time prodding students to think about what makes an appealing essay topic, and why certain opening sentences are more effective than others.

Teacher Effectiveness and Satisfaction

This re-allocation of teacher time gets at a third major benefit of blended learning: a far more satisfying teaching experience, at least for teachers who embrace the use of technology to make themselves more productive. The definition of a tool is something that makes it easier to accomplish a given task. In a blended-learning regime, technology makes teaching easier and more efficient. If you go into the teaching profession to make a difference in children’s lives, which presumably most teachers do, having effective tools can make all the difference in the world.

There will certainly be growing pains as teachers learn to operate in new ways. On the first day of a pilot blended-learning math program in a fourth-grade classroom at Visitation Catholic school in North Philadelphia, teacher Mary Anne Corcoran is working hard to figure this out. With 16 kids in her class, half are supposed to be working on DreamBox, and half with her. But with only five machines up and running, she has the class in three groups: computer, teacher, and individual work.

The DreamBox kids wear headsets; the ones doing problem sets in their workbooks do not. Corcoran starts out in usual teacher mode, up at the chalkboard using her teacher voice to instruct her small group. Naturally, the individual-work group finds their teacher far more interesting than their pencils and workbooks, and they start watching her instead. Seeing this, she swiftly abandons the chalkboard, and moves closer to her instructional group, speaking much more softly.

In normal classes, there is little time for individual interaction, as the teacher tries to keep everyone focused on the same material. Keeping the attention of a group of children sometimes requires heroic effort, or at least a magician’s bag of tricks: call and response, moving around the classroom, cold calling. But why should running a classroom be limited by the demands of crowd control? Teaching is about pedagogy, not herding cattle.

Ideally, blended learning will allow a smaller number of truly wonderful teachers to preside over more students. Bryan Hassel of education consultancy Public Impact applies the label “3X” to these teachers for their ability to have three times the effect of normal sorts. This isn’t an exaggeration; studies have found that repeated exposure to excellent teachers can help students from lower socioeconomic backgrounds actually close the achievement gap. The problem is that if you insist on having only 20 to 25 students in a class, there just aren’t enough 3X teachers to go around.


Summit students

With blended learning, the field won’t have to rely on so many mediocre teachers, because less-complicated (if sometimes onerous) work will be outsourced to machines and teacher’s aides. These aide positions, freed of the heavy credentialing demands of traditional teaching, will appeal to a different group, drawn in by aspects like the part-time work week, which may appeal to parents of young children. And principals could maintain their teams at a smaller size, focusing more resources on the high performers who remain. With good tools and higher pay, these remaining teachers are likely to be more satisfied.

Some smart philanthropy is betting that increased teacher satisfaction will be the catalyst that truly spreads blended learning in schools. Money spent introducing teachers to blended learning will thus be money well spent, goes this thinking. CityBridge Foundation, which focuses on Washington, D.C., has made introducing teachers to blended learning a key part of its giving strategy. After a carefully studied pilot project at D.C. Scholars Stanton Elementary School in summer of 2012, CityBridge launched the New Schools Education Innovation Fellowship later that year. The foundation selects a dozen teachers who meet for one day every month to hear from blended-learning experts and see technology demonstrations. They travel to blended-learning schools in other states over their breaks. During the summer, they each test a blended-learning pilot program. The fellows later implement their pilots in their own schools.

“We believe creating a cohort of teachers and helping them to become entrepreneurs, to take risks, and to innovate is going to set our city up to really embrace blended learning in the right ways,” says Emily Silk, who handles blended learning for CityBridge. The foundation hopes to select new fellows annually to build up the ranks of ambassadors for blended learning.

“It’s a new thing to learn how to teach this way,” says Summit math teacher Jesse Roe, who taught for many years using technology, but not in a blended model. “But that makes it interesting. For me, the results prove this is the way to go.” In its first year, Summit’s new blended-learning model was certainly “a lot more stressful and a lot more work.” On the other hand, “it was more enjoyable,” says this math teacher. And “we feel it’s better for our kids.”

Cost Control

Most education money comes from state and local governments. Many states, counties, and cities find themselves in dire fiscal straits today, and the public mood is turning toward austerity. One of the major attractions of blended learning is the possibility of eventually getting better results without extra spending, and perhaps even getting better results while spending less than at present.

Cost control is not easy to achieve in education. While most goods have gotten cheaper over time, education has seen very few productivity gains, and over the last generation it has, in fact, become a lot more expensive per unit of output. There is, however, no reason that education needs to cost as much as Americans are now paying. Our costs per pupil are much higher than in South Korea, Finland, and other high-performing countries that also provide lots of teachers, modern buildings, and ample books and amenities. Decades of studies by scholars like Stanford economist Eric Hanushek have shown that Americans have gotten very little from the gusher of money that has been directed into K–12 education over the last generation.

With many states and the federal government now in fiscal crisis amid record spending, K–12 spending in the U.S. will have to become more productive. The tantalizing prospect of blended learning is that it could help states and districts do just that. “I don’t frame it as lower cost,” says Ethan Gray, director of the Cities for Education Entrepreneurship Trust, a network of city-focused foundations and mayors’ offices that support education reform. “I frame it as a different business model. The easiest way to put it is it makes good teaching less expensive.”

Summit Public Schools head Diane Tavenner identifies the obvious starting point. “If a computer can replace a teacher, it should. We need to identify every single place where a computer can legitimately replace a teacher and replace them there.”

One place this is already starting to happen is in higher education. Virginia Tech, for instance, now teaches freshman math in a giant computer lab called the “Math Emporium.” It is housed in a space that used to be a discount department store. Picture hundreds of students, each sitting at an Apple computer with a red plastic cup next to it. Students work through the freshman math curriculum online. If they get stuck and require help from a human being, they place the red cup on top of their monitor. Circulating teachers (who are not professors) come alongside to answer questions and help students stay on track.

In many ways, the computers provide a surer product than you’d get from a grad student who is teaching freshman math to earn his stipend. Traditional lectures give students little chance to actually do math. And lecture classes don’t help if a student has a crucial gap in his background knowledge. The result, according to the Washington Post, is that Virginia Tech students now pass math classes at a higher rate than they did previously, at one-third less cost.

Some K–12 blended schools are also starting to see productivity gains. Rocketship schools generate 15 percent margins on their state’s standard per-pupil allotments—which they then plow into opening new schools, training teachers, and other priorities. Carpe Diem is likewise financially sustainable because of its 75-to-1 pupil-to-teacher ratio. Educators at KIPP Empower Academy in Los Angeles originally adopted a blended model to make the most of the increased class sizes that California’s student funding cuts necessitated. They are now getting impressive results with official class sizes of about 40 students (who rotate in sub-groups among stations).

Purchasing computers and software to start a blended-learning school is expensive. Many foundations have paid for planning grants and for consultants to help schools select software and design their programs. This start-up capital has been extremely helpful for these schools, though it raises concerns about scalability.

The good news is that planning becomes less complicated with each passing year. The Learning Accelerator, an organization started by investor Joe Wolf and funded in part by the NewSchools Venture Fund, works to fill existing gaps in the ecosystem of education technology. One project will create a partially standardized package for starting a blended-learning school. “We’ll have arrived at some basic standards, some basic protocols on what a good implementation looks like,” Wolf says. That way, consulting work can be “a $30,000 project rather than a $300,000 project.”

On the hardware front, some schools might have a “BYOD” policy: bring your own device. College students bring laptops to class; in some places K–12 students are also expected to acquire a computer, sometimes with help from revolving loan funds or group-discount options. Philanthropists could make more targeted investments for families who can’t afford technology, versus the majority who (given the market penetration of smartphones and laptops) clearly can. Meanwhile, tablet and laptop prices are falling rapidly; at $200, purchasing a new machine for each student can be cost-competitive with the usual budget for textbooks and photocopying.

There are still reasons why blended-learning schools might not see immediate productivity gains. Software costs money, as does teacher time as educators figure out new ways of working. If schools need to employ a technology specialist, this eats up some of the savings gained by bigger class sizes. And if you don’t adopt bigger class sizes, you won’t see any cost savings at all.

There is, however, good reason to expect that blended learning could eventually help schools function well even on reduced per-pupil allocations. If they can function well at lower basic costs, they’ll be able to afford the extras that attract parents and students: art, music, field trips, and even nicer school buildings to match all their brand-new technology.

For the first time in a while, funders say, there’s a sense within the education-reform movement that we’re about to have tools that will enable good teachers to become great. Philanthropy can play a big role in making that happen—in identifying promising ideas, nurturing their growth with smart investments, and spreading them as broadly as possible. If it does, America will come closer than it ever has to a schooling system where all children get the education they deserve.

Laura Vanderkam is the author of The Philanthropy Roundtable’s new guidebook Blended Learning: A Wise Giver’s Guide to Supporting Tech-assisted Teaching, from which this article is adapted.

Let’s Keep in Touch

Our Values-Based Giving Newsletter helps philanthropists and charitable organizations apply their values to their giving and follow the best practices for success.


This field is for validation purposes and should be left unchanged.