In “Straight Talk with Rick and Jal,” Harvard University’s Jal Mehta and I examine some of the reforms and enthusiasms that permeate education. In a field full of buzzwords and jargon, our goal is simple: Tell the truth, in plain English, about what’s being proposed and what it might mean for students, teachers, and parents. We may be wrong and will frequently disagree, but we’ll try to be candid and ensure that you don’t need a Ph.D. in eduspeak to understand us.

Today’s topic is “high-dosage” tutoring.

—Rick

Rick: A lot of folks are enormously bullish on the power of tutoring, especially after the pandemic. And look, I believe that tutoring can be a terrific thing. As Bror Saxberg and I observed a decade ago, “One-to-one tutoring with a good tutor is about the best way we know to provide intense instruction, real-time customized assessment, and intensive, personalized practice.” And there’s solid research to support both the intuition that tutoring helps and also that “high-dosage” tutoring of 90 minutes a week yields outsized benefits.

Now, here’s the “but”: Whatever its potential may be, tutoring has historically been expensive and logistically challenging to pursue at any kind of scale. That’s why individual families usually purchase it themselves from paid tutors, local voluntary programs, or online vendors. Schools have rarely offered anything more than desultory, mostly cosmetic offerings. And that’s because it’s really tough to do more.

A number of years ago, the Houston school district launched Apollo 20, an ambitious tutoring experiment for students in grades 5 and 9 in targeted middle and high schools. The effort, for two grade levels of students in a limited number of schools, was backed by millions in dedicated funding and intensive hands-on support from a Harvard research team. Yet, even with all these advantages, recruiting, training, and retaining enough part-time tutors proved daunting. The idea was a good one, but, ultimately, just too difficult to operate at scale.

It seems to me that the biggest challenge with tutoring is making it practical, accessible, and consistently good. And I fear that the “high-dosage” label is unhelpful here. I know it’s supposed to point to 90 minutes a week, but it’s turned into one of those labels, like “smart assessment,” that means pretty much anything. In fact, it seems like school and system leaders have started reflexively describing all tutoring as “high-dosage” just to be safe.

So, I guess I want to see what happens in practice. Can tutoring programs recruit, train, and then retain enough good tutors? Will AI-in-your-pocket (like Khan Academy’s new “Khanmigo” tutoring program) really change everything? And can schools create routines that make it possible for “high-dosage” tutoring to be more than an aspiration?

Jal: That all sounds right. Matt Kraft, who has both done the research on tutoring and has offered a blueprint for how we might scale it, suggests that it become a regular part of the school day rather than an after-school add-on, precisely because that can create the kind of routinization that sustainable programs need.

Tutoring is easier than teaching—many fewer students, shorter time periods, more focused goals—but you still have to motivate students and you need some subject-specific content and pedagogical knowledge to help struggling students. Just because you know how to read doesn’t mean you know how to teach reading, which is why training is critical. Achieving quality at scale will not be easy, and a lot will depend on: 1) the nature of the training for the tutors; 2) the availability of strong, curriculum-aligned materials; and 3) local leadership to pick the tutors and oversee the whole enterprise. Given that, it seems that rather than simply spending huge amounts of ESSER funds on tutoring, districts and schools should only move forward if the above conditions are met.

One thing that I find intriguing about tutoring is that it is structurally innovative but pedagogically and philosophically conservative. If incorporated into regular school, it would be a structural innovation in that it would move part of the day away from the familiar 1-teacher-to-25-student model and replace it with individual and small-group instruction. It would also be structurally innovative in that it would greatly diversify the types of people doing the teaching. But all of this effort would be devoted toward the existing goals and assessments of schools—diversifying the adults but giving them the same jobs. What if we extended the idea of connecting students with a wider variety of older mentors but took advantage of the fact that this village might have many more things to teach than what you get in regular school?

Rick: I really like this “structurally innovative, pedagogically conservative” point. You’re right that most tutoring is focused on mastering traditional content. And I’m wholly good with that. But, as you note, tutoring can help do a lot more.

For instance, Julia Freeland Fisher and Daniel Fisher explore in Who You Know that virtual tutoring-like tools can help schools expand students’ access to relationships that might otherwise be out of reach. In a world full of active retirees, remote workers, and the self-employed, it’s a lot easier to find adults with the flexibility to engage and the time and interest to serve as mentors. And they note, just as Bob Putnam has documented in Our Kids, that these kinds of “loose” ties matter a ton in terms of college-going and getting ahead.

I’m also wondering, though, about the remarkable enthusiasm regarding the power of AI-enabled tools to upend the tutoring paradigm. Robin Lake, head of the Arizona State-based CRPE outfit, who is thoughtful and pretty measured, summarized her takeaways from a big ed. conference in May by declaring, “In a matter of weeks or months, [AI tools] are going to be your kid’s tutor, your teacher’s assistant and your family’s homework helper.”

I’m intrigued by the possibilities of AI and the power of these new tools, and we’ve seen that ed. tech used at home has consistently been far more transformative than the stuff sold to schools. But a century or more of overhyped ed. tech has left me pretty wary. I find it a lot easier to see students using AI to get answers than to master concepts. After all, it’s safe to say that advances in computing have done a lot more to fuel gaming, social media, and shortcut-taking than to promote learning. I think there’s pretty compelling evidence that this has been bad for student social development and mental health. While I appreciate the potential of AI, it feels like talk of transformation is pretty premature—and I can already picture the hash that “early adopters” could make of all this, with the best of intentions.

Jal: I thought we were supposed to be debating! I agree with that as well. Education is fundamentally a human enterprise. Young people, particularly but not exclusively those who are struggling in school, need people to believe in them, people to reassure them when they struggle, and people to celebrate their successes. Tools could conceivably help, but at their best, they are just one piece of the larger puzzle. As Justin Reich shows in his masterful survey of education technology over the past century, there is a consistent Matthew Effect to new technologies: Those who possess more academic, cultural, and social capital tend to benefit the most from new tools and be given opportunities to use them in the most advantageous and agentic ways, while those on the shorter end of the stick tend to be asked to use technology in ways that make them more passive consumers than active users. For example, while kids in private schools are exploring the possibilities of ChatGPT (alongside their teachers), students in public school districts—particularly higher-poverty ones—are banning ChatGPT from schools. We should know by now that technology isn’t going to save us, although it can be one part of the equation.

To return to human tutoring, I think it is promising precisely because it provides individual or small-group human connection, particularly for struggling students. You could imagine a world where a student spent much of the regular school day trying to conceal their weaknesses to avoid embarrassment but where working privately, one-on-one, with a trusted mentor could lead to real progress. I coach youth sports and I find a lot of value to having myself or another coach work with kids individually outside of regular practices, because it allows for skill development in a private space where it is really OK to try things, fail, and slowly get better. But, and this is a huge but, this is all dependent on the tutor actually knowing something about the domain and how people learn in the domain. This again brings us back to the coaching and the selection of the tutors. As with most things, tutoring can be good but only if we really think through the conditions needed to do it well.