This interview focuses on a chapter from A Nation at Risk +40, a report by the Hoover Education Success Initiative (HESI) that looks back at the birth, struggles, and future of the modern school reform movement. (Download the publication here.)
Tom Vander Ark is a longtime education consultant and CEO of Getting Smart. He was also the first executive director of education at the Bill and Melinda Gates Foundation. His chapter deals with the impact of rolling out computers, Internet access, and education technology in US schools. He spoke with Chris Herhalt about the future of “edtech.”
Chris Herhalt: You’ve pointed out that US school districts spend $44 billion a year on electronic devices, edtech software licenses, the Internet access. You add, “There’s been limited aggregate improvement in basic skills as a result of this investment.”
Aside from the COVID-era losses in learning, why is this?
Tom Vander Ark: That’s a big question that tortured me in writing this paper, since I’m a tech optimist and have been something of an edtech leader for thirty years and have been pretty optimistic that it was going to prove everything, everywhere. The fact that it doesn’t appear to have had much benefit was difficult to live with for the months of putting this together. The answers are complicated, but the two topline answers are: one, it probably made more difference than we know because it improved a lot of skills that we’re not very good at measuring, and two, it wasn’t deployed well.
Over the forty years we’ve lived in a standards movement, we’ve become fixated on measuring proficiency at grade level in reading and math. Those were the only two things that mattered. We didn’t see any appreciable growth—or, during the pandemic, we gave back any growth that we’d achieved over the prior twenty years. But I would say that in the past ten years, certainly the past five, most American school systems have begun to update their learning goals: we finally started paying attention to twenty-first-century skills, to focus on creativity, critical thinking, and collaboration. And I suspect that technology and blended learning probably helped improve those things. So that’s the first answer.
The second answer to why the tech didn’t work is bad deployments. If we look narrowly at adaptive learning applications, which seemed to be the category with the highest promise for improving performance, each one of the providers—this would be in both reading and math—can show evidence that when the recommended dosage is achieved, the results are pretty good. I think it’s fair to say that you could go to any of those adaptive learning providers—Curriculum Associates and DreamBox and Istation, which has now merged with Amira, IXL—and where they had strong deployments and most kids were getting the effective dosage, their results were pretty good.
The access to technology and then learning apps was typically adopted in a rather chaotic fashion. A bunch of it came in bottom up: from teachers’ scrounging and schools that had a little bit of money left over at the end of the year. Some computers showed up, they may or may not have showed up at the same time as some learning applications, they almost never came with enough training, and so the mere adoption of technology was super chaotic. As a result, very few places created new learning models and had high-fidelity use.
I noted in the paper that the groups of schools that seem to have the biggest measured achievement gains were the CMOs, the charter management organizations, as noted by Macke Raymond and her CREDO study. The third CREDO study showed CMOs were quite effective compared to traditional public schools. And I think the main reason is that they all developed a coherent learning model, their schools all exercise high fidelity to that learning model, and the learning model incorporated effective uses of education technology. So those are the first two big answers. It probably was better than we measured, and I think effective deployment made the biggest difference.
Let me add an interesting reflection on why tech deployment was so bad over the past thirty years. I think it’s a balance sheet problem. I’m a finance guy from the private sector and taught finance in a couple of MBA schools; I was VP of finance for a public company before I was a school superintendent. In the private sector, when a company deploys new technology, it’s a capital expenditure. You can plan for and execute a well-supported deployment of new technology and combine it effectively with talent development, so that the new technology shows up when the workforce is trained and ready to use it. Because you can do this off your balance sheet, you raise money or debt, or use retained earnings, in a sequenced, thoughtful way.
School systems, charters, and districts don’t really have a balance sheet. School districts can run a bond to borrow money to build a building, and operating funding is a weird, complicated mixture of local, state, and federal funding—but there’s no facility to make investments in durable assets like technology. In short, schools don’t have the financial or organizational wherewithal to deploy technology in a thoughtful or effective way. This is a core governance problem. We haven’t created school systems with the financial ability to manage what is now a permanent part of the infrastructure.
Herhalt: At one point in your chapter, you cite a study that says, “Up to 67 percent of edtech applications acquired by school districts either go unused or are incredibly lightly used.” What’s going on there?
Vander Ark: I talked about this chaotic combination of bottom-up adoption by teacher and then sloppy, poorly planned and -executed top-down implementation. When you mash those two together, you get a lot of extra stuff. You had curriculum directors going crazy, buying up apps and adding subscriptions that nobody ever accessed. And then you had teachers doing it individually, and schools doing it individually, and you had departments like special ed grabbing applications. It was probably exacerbated by a lot of freemium stuff, where you could use it for free and then it would roll into a subscription. It’s unbelievable how many subscriptions went unused.
There’s probably a dose of that in the corporate world, but I’m sure it’s much more prevalent in education. Remember that when edtech started, we had a tradition, particularly at the secondary level, of teachers teaching whatever they wanted. Teachers were basically individual practitioners and school boards could adopt textbooks, but they were suggested. And then as technology rolled in, this created an explosion—we call it unbundled learning—where teachers just started using stuff.
Herhalt: As a parent, I’ve twice been moved to ask during parent-teacher conferences, “What else can my daughter be doing? What else can be offered to her in terms of enrichment, extra work, and so on?” The only thing I’ve been offered is, “Well, we have these edtech applications. She can go and do that.” I don’t think that was anyone’s intention when edtech entered the classroom. What do you think of it being framed this way?
Vander Ark: That’s an interesting question. The way you frame this is exactly how the Los Angeles Unified School District framed Ed, that chatbot that blew up the company behind it that just collapsed. It did two things. It was a chatbot for parents to figure out what was for lunch and where to get the bus. And then it was, “Oh, and your daughter can do this exercise in IXL.” It pointed kids at grade-level learning exercises. We’ve automated that idea of a narrow set of edtech applications that are the extension of school. I guess, like you, I find that unfortunate.
I’m thinking about the hundreds of school networks that I really appreciate which are much more project-based, more progressive, that are still probably not much better at answering your question of “what else?” They might recommend more reading, they might recommend some project-based ideas, but I still don’t think they’re a lot more skilled than traditional schools. Only a handful of schools have really made the shift to competency-based learning, where they’re really clear about what kids need to know and be able to do and how they will demonstrate their learning, and where kids get that and where they can own their progress. And in those schools, kids can take a lot of responsibility because they know what they need to do to move to the next level. They can really answer this question of “what’s next?” for themselves. I would say very, very few systems have achieved that level of maturity and sophistication.
Herhalt: Is there a role for the states or the federal government, maybe not to control how this is done, but even just to compile some best practices and say, “Here’s what works and what doesn’t”?
Vander Ark: Yes. The best answer to your question is that we should blow up the system we have and develop one that looks like a modern school system. This comes back to the governance problem. We’ve inherited a nightmarish governance structure that’s very, very poorly suited for doing the work. Between local educational agencies [LEAs], there are probably 13,500 districts, and 5,500 charter LEAs, so together you have 19,000 leadership teams trying to figure this out. It is way too complicated.
I think there are probably twenty or thirty, not more than fifty, school-system leadership teams sophisticated enough to design a coherent tech-enabled learning model, to develop a tech stack, and to deploy it and then update it dynamically. That’s fifty fully equipped teams, while 19,000 teams are trying to do this. I’d like to blow up the whole system and flip to networked schools under performance contracts. And I’d like to see twenty or thirty dominant learning models operating around the country, and then invite schools to join the networks that share that learning model.
Short of that, one layer of the system I didn’t talk about in my paper is the intermediate unit. They show up in almost every state. And a few of them are good; most of them are not. But that would be another place where you could manage technology or support technology. Deployment would be at that intermediate unit level. And I think a governance push to at least get learning models and tech models, either at an intermediate unit or in networks of like-minded schools, would be one approach.
But you mentioned advice from states and the federal government. I want more than advice. I want to try to move more schools into platform networks. In a platform network, you have a learning model and a school model on a digital platform combined with professional learning experience, like a talent development system in a network. That can be either a place-based network or a thematic network that operates nationally. Like NAF: they have 620 career academies with a pathway strategy. They have shared curriculum and it’s all on a common platform. They have a teacher development system. It’s national in scope. I think we need more of those. There are a handful of districts that have achieved this coherent learning model, aligned with a strong tech stack and a talent development system. So, you can’t just provide advice on how to deploy technology, you have to actually build a full model and support its deployment.
Herhalt: Among all these developments of edtech applications, we’ve also seen the proliferation of kids with smartphones. Certain jurisdictions are imposing blanket bans on them in the classroom. Where do you fall on this?
Vander Ark: I’m smiling because they buttonholed me at ISTE [International Society for Technology in Education] when I was getting one of their ISTE Impact Awards for leadership in edtech. I said, “This is really ironic, when I’m sitting on a stage and getting this edtech award, that after spending thirty years advocating for access to technology, I’m going to say I think we should ban cell phones.” It’s a blunt solution to a real, nasty problem. And it’s not the phone so much; it’s the social-media apps that have become so problematic. Jonathan Haidt talks about how it’s been a collective-action problem, that individual kids are using social media in a problematic way but it’s the joint uses that are so dangerous. So, it really does demand a collective-action solution.
Much better than outright bans would be shared-use guidelines, so that there’s a community agreement about how and when we’re going to use certain tools. You also could extend this to artificial intelligence. You could have a school community come to a nuanced view of these tools and as a community would police that set of uses. That would be a lot better than a ban. But I think it’s really hard to get to nuance in a community. So, if I had to pick ban or no ban, I guess I’d go along with a ban. I’d want to term-limit it for two years, and then revisit the policy.
