How hard is tutoring to get right?

Today we’re wrapping up our deep dive into the topic of tutoring with comments from Mike Goldstein, Matt Kraft, Harry Patrinos, Dan Goldhaber, and more; reading two reviews of James Traub’s powerful new book on civic education (by Mark Lilla and Rick Kahlenberg); and rounding up other recent posts by Anna Egalite, Karen Vaites, Kerry McDonald, and Ben Austin, among others.

Sign up to receive this newsletter in your inbox, usually on Tuesday and Friday mornings. SCHOOLED is free, but a few linked articles may be paywalled by other publications.

One of my main concerns with tutoring as “the next big thing” is that, to my eyes, it looks too hard to implement effectively and at scale. Not in terms of the politics; as Kevin Huffman wrote last week, tutoring is super popular, both with parents and with educators. And why wouldn’t it be? Everyone likes extra services for kids who are behind, especially personalized ones, and especially if money is no object.

No, my concern is that there’s “a narrow path to getting it right,” as Mike Goldstein and Bowen Paulle wrote several years ago in a report for Fordham. The tough challenges include finding, training, and scheduling the tutors; matching the instruction to what students are learning in class; and keeping the initiative a priority among changing district circumstances. And the greater the scale, the larger those challenges loom.

But is tutoring any harder than any other school reform, including those I favor, like dramatically raising salaries in high-poverty schools? Mike himself is doubtful:

The Narrow Path applies to all the policies you are proposing as well. There are so many counter-examples to your preferred strategies. In some places teacher salary has hit $110k/year or whatever with zero gains to kids. The literature on HQIM or “good curriculum”—same (look at Tom Kane showing how math curricula made zero difference). The literature on “good” teaching training is WAY worse than the choppy results of tutoring; tutoring sometimes survives the RCT, training almost NEVER does. I know you know this because Fordham covers all that. In Liz Cohen’s book, she stipulates that you need good execution. And I do think the research shows if you have good execution, then high-impact tutoring eclipses other strategies—perhaps because at its core, tutoring one struggling kid at a time is doable by a mere mortal, while teaching 20 strugglers at a time (i.e., classroom teaching) is just so much harder to achieve.

Matthew Kraft, meanwhile, cautions patience:

My take is while the post-pandemic tutoring evidence has been less positive than hoped, that is in part because our expectations were inflated and our evaluations were premature. These programs take time to get right at scale.

I think we got the order wrong in our understandable attempt to help as many kids as we could as fast as possible. We went for scale and are now trying to get to quality. We need to focus on quality and then scale horizontally (lots of small programs) instead of vertically (expanding program size).

But also count me on team “let’s make targeted investments in teacher compensation and provide educators with great support and instructional materials.”

Matt Barnum thinks we’re being overly glum about the impact of tutoring:

I admittedly have not dug into the research super closely or read Liz’s book yet (though I plan to). But I don’t look at this graph and think all the money on tutoring has been “wasted.” I think we’ve run a bunch of underpowered studies but in aggregate they clearly lean positive.

I mean these effects are not quite as big as we might have hoped but nothing to sneeze at.

So where did the idea that small group tutoring failed and was a waste of money come from? Or is it just that it fell short of the highest effect sizes from prior literature?

Harry Patrinos is even more enthusiastic, based on his read of the international evidence:

Tutoring is challenging and can waste resources when poorly implemented. But mounting evidence confirms its effectiveness. Done right, high-dosage tutoring substantially improves outcomes for struggling students.

The obstacle? Cost. One-on-one and small-group tutoring can be prohibitively expensive, even when cost-effective, unless education budgets grow significantly.

Online tutoring with live instructors offers a powerful alternative. Multiple experiments, accelerated by Covid-19 school closures, demonstrate that remote tutoring works—and works well.

The evidence is compelling. In Italy during lockdowns, university students volunteered to tutor disadvantaged middle schoolers online for three to six hours weekly. A recent study shows striking results: Math scores increased by 0.22 standard deviations in 2020 and 0.20 in 2022, with larger gains for students receiving more hours.

Similar findings emerge from Kenya, Spain, and the United States. American studies confirm online tutoring’s cost-effectiveness, with learning gains between 0.05 and 0.23 standard deviations, the upper bound approaching a full year of learning.

Perhaps most remarkably, online tutoring works even amid war. Our study in Ukraine tracked nearly 10,000 students across all regions between early 2023 and mid-2024. Through three randomized experiments, students in grades 5–10 received three hours weekly of small group tutoring in math and Ukrainian language for six weeks.

Despite displacement, blackouts, and air raid sirens, the program delivered learning gains of 0.49 SD in math and 0.40 SD in language, while meaningfully reducing stress.

These results suggest that online tutoring and blended learning may help address Bloom’s two sigma problem—the challenge of replicating the effectiveness of one-on-one instruction at scale. While that may be ambitious, the lesson for policymakers is clear: Invest in high-dosage tutoring and leverage online delivery to expand reach while controlling costs.

Technology has made tutoring affordable, scalable, and effective. The question is no longer whether it works, but how quickly we deploy it.

Still, Dan Goldhaber agrees that execution is everything:

Tutoring is another one of those issues where we have one word to describe what is effectively many different types of initiatives and an intervention in which implementation matters.

I don’t think we should set this up as an either/or. I think tutoring can be tremendously effective if well targeted, if kids aren’t missing foundational lessons from regular teachers (sometimes happens!), if it’s not voluntary, and probably a lot of other “ifs,” too. Unfortunately, the level of coordination needed to do tutoring well (at least during a pandemic/flood of ESSER money) may make the average effect of tutoring less than the average of summer school.

Tutoring reminds me of teacher evaluation/compensation reform in that it can have quite large beneficial effects on students, but only if done right.

And doing it right is particularly important—and challenging—when it comes to English language arts, argues Thibaut Delloue:

I wanted to share my thoughts on tutoring, as I have personal experience with teaching classes and tutoring in literacy. I agree with you that tutoring, as it’s currently conceived, is a tremendous waste of resources. The underlying problem, however, is the same as the problem that ails academics in general in American public schools. David Steiner put it best: “Three of the major pillars of our education system—how we prepare teachers, what we test, and what they teach—embody industries that exist in their own bubble”—though I would include academic standards there, too.

Literacy tutoring rarely works because tutoring programs have their own curricula and standards. If I plop a kid into a tutoring space for 45 minutes and let them learn whatever the tutoring program has dreamt up, it’s not going to make much impact on their progress on a whole other system of state standards, whichever private curriculum the school has adopted, and whichever private assessments the school has purchased (NWEA, iReady, etc). There is no meaningful way to know if tutoring is effective if it’s completely divorced from what kids do in their regular classes, a problem that is mirrored in schools in general: It’s hard to gauge the effectiveness of instruction when all the major pillars of our academic system are made by different entities and only tenuously “aligned.”

The underlying issue here is that background knowledge is integral to and inherently part of learning—not just for “reading class,” for all academic subjects. This has been proven over and over in research. Most other countries have academic systems that are conceived through the state alone, so aspects like standards, curriculum, assessment, and teacher prep are perfectly in sync. That means, unlike in America, kids are learning a coherent system of knowledge.

So, tutoring will be actually effective if the content kids are reading is directly tied to what they are learning in school. Instead, many tutoring programs emphasize generalized “reading skills” like “finding the main idea” or “learning a sequence of events.” We know now that this is not effective; see Daniel Willingham on that issue. Yet so many of our public schools still view academics as content-neutral. So unless we fix how we understand academics in the first place—which is most of what schools teach—then tutoring will not be impactful either. Learning—and especially literacy—is at its core an individual process that happens in your brain. One-on-one instruction, if done right, CAN of course be effective. But first the education establishment needs to shake off decades of being misled by schools of education—which are run not by educators but sociologists—and wake up to how human beings actually learn.

Finally, my colleague David Griffith homes in on one of the most important issues: that we can’t effectively tutor everyone, but we can make targeted tutoring work:

I share many of your concerns about tutoring. But I think it deserves a fair hearing. And I’m not sure it got one in the post-pandemic era, given the insane time-pressure that districts were facing and the massive scale of some of the programs.

No, it doesn’t make sense to tutor every kid who is behind, especially in places where most kids are behind (and classroom instruction may be meeting them where they are). But what if the choice is between “more” paraprofessional teaching assistants (or other non-teachers) and a modest corps of tutors hired and trained specifically to help the lowest 5 percent of students learn to decode?

To me, the latter is worth considering, provided we’re willing to take it slower than we have been and commit to doing it right.

---

• James Traub has a new book out, The Cradle of Citizenship: How Schools Can Help Save Our Democracy. SCHOOLED discussed an excerpt last week about the resurgence of classical education, about which Jim was originally skeptical but came to like. The New York Times assigned Mark Lilla to write a review of Jim’s book, and it’s well worth your time. Mark writes:

  A few years ago, Traub, a former contributing writer at The New York Times Magazine, had the excellent idea of heading out to visit a number of publicly funded schools across the country to see how teachers and school districts are managing [to cope with the culture wars roiling America]. The good news: Most are trying their best…

  As in most organizations, the people on the ground were just trying to get the job done while dodging directives raining down from above.

  But that’s where the good news ends. The real crisis of civics education, Traub discovered, is not that students are learning about 1619 rather than 1776, or the reverse. It is that so many are learning nothing at all. And here he lays responsibility at the feet of the American educational establishment, which has, in the words of one scholar, been turning the “meat of academic subjects into meatloaf.” (Emphasis mine)

  Yes, yes, yes. Such arguments will resonate with anyone who’s been reading Don Hirsch for the past three decades. Mark concludes:

  Traub does not offer the classical model as a panacea for American civic education. But he is certainly right that whatever model is used should impart “a solid foundation of linguistic skills, historical knowledge and habits of reflection” and offer “an alternative to the consumerist, vocational, instrumental culture of today’s public education.”

• Richard Kahlenberg has strong views about civic education, too, from his perch as the director of PPI’s American Identity Project, and penned the Washington Monthly review of Jim’s book, also worth reading. Rick writes:

  Traub is dismayed that schools in both red and blue states often get things wrong on three major issues.

  First, in the perennial debate over whether students should learn by reading or doing, and pursue knowledge or skills, the obvious answer is that they should do some of both…

  
  Second, in the legendary debates over how to teach American history, Traub is troubled that some red states take a jingoistic approach that soft-pedals America’s sins, while some blue states teach versions of American history that suggest oppression is the overriding theme…

  
  Third, in the important debates over how our schools should grapple with the contemporary issues of racism, Traub sees troubling signs on both extremes.

  Both Rick and Jim are optimistic, though, that a commonsense middle ground can be found and point to support for the Educating for American Democracy roadmap as one encouraging sign. “As checks and balances erode,” Rick writes, “the schools may be our last hope.”

• If schools aren’t teaching “historical knowledge” and the like, what are they teaching? In some cases, it’s “social and emotional learning,” or SEL, which advocates have argued is not only necessary in its own right (we do want kindergarteners to learn to share and control their impulses, after all), but can boost test scores, as well. The evidence around the latter point, however, is super weak, Anna Egalite shows in a devasting research review for Education Next. Anna writes:

  Outlets like EdWeek and NPR report that SEL can raise achievement by 4 to 8 percentile points, citing new data as “clear evidence” that SEL programs lead to better grades and test performance. These claims all trace back to a new meta-analysis from USC and Yale that has already shaped national coverage. But before policymakers take these estimates at face value, it’s worth looking closely at how the study was built.

  Take a close look Anna did, and she found multiple problems with the meta-analysis that raise fundamental questions about its conclusions. Among the issues: Many of the included studies had tiny samples, were about programs overseas, included the same data multiple times, and conflated wildly different measures together. “The idea that ‘cognition and emotion are inextricably tied together’ (as NPR quoted one of the authors as saying) is neither new nor controversial,” Anna writes. “Few educators would disagree with that conclusion. But demonstrating a broad relationship between emotions and learning is not the same as providing rigorous evidence that any given SEL program will raise students’ test scores.”

• It was too late to include on her Christmas list, but for the New Year, Karen Vaites wants a national curriculum database (don’t we all?). Karen writes:

  Imagine a world where we could run regression analyses to see the impact of curriculum change on district performance. Or we could compare the aggregate performance of districts using curriculum with and without whole books, or basals vs. knowledge-building materials. Imagine being able to see which programs are associated with district overperformance—and underperformance.

  It’s a beautiful vision, though I’m skeptical that Linda McMahon is going to make it a reality, as Karen would like. It’s also a harder problem than it might sound. Morgan Polikoff has spent years trying to collect data about the adoption of specific instructional materials, so I asked him for his thoughts. Morgan replied:

  Most of what I said here still applies. That said, if there was a centralized effort that was required in some way (not sure under what authority), that’s probably the single best way to do it. Still there would be lots of challenges like a) what about different textbook versions, b) what about places where they use multiple materials (for different grades or across schools), c) what about timing, d) what about supplementals, e) what about open-access, f) what about teachers’ use, etc.

• Last week, SCHOOLED noted Jorge Elorza’s enthusiasm for potential 2028 presidential candidate Rahm Emmanuel, given his embrace of education reform. This week it’s Ben Austin’s turn, in his case to laud Pennsylvania Governor Josh Shapiro:

  Shapiro’s support for school choice puts him at odds ideologically with Democratic stars like New York City Mayor Zohran Mamdani and Rep. Alexandria Ocasio-Cortez (D-N.Y.). Yet he is the most popular politician in the most important swing state in America. Given the stakes for American democracy in 2028, that’s a data point Democrats ignore at our collective peril.

  Neither Jorge nor Ben are expressing their support for a particular candidate, but are instead pleading for the Democratic Party to embrace reform, which, according to Ben, signals “a willingness to confront internal orthodoxies when they collide with the public’s demand for results.”

• Goldy Brown and Christos A. Makridis are happy that the Trump administration moved career-tech programs from Education to Labor. This shift “acknowledges that education policy cannot substitute for labor market policy when the objective is work,” the authors write. Which makes sense, as do Goldy’s and Christos’s other points about CTE done right. But in terms of moving federal responsibilities around, CTE to Labor is the easiest case. Now I’d like to see someone make a similar argument for sending OESE to Labor—given that many of its programs have objectives other than work. I think it’s hard to do.

• Kerry McDonald loves that ESA programs are allowing students to switch schools mid-year if their current schools aren’t working out. That’s especially true for micro-schoolers. Kerry writes:

  Unlike many traditional private schools—which typically have set admissions and enrollment cycles, lengthy application processes and hefty tuition price tags—today’s emerging schooling models are usually low-cost, flexible, and highly personalized. They often have rolling admissions, with students able to enroll throughout the year.

  That sort of responsiveness is great for families—though as we’ve learned from the charter schools movement, policy needs to be responsive, as well, shifting public funding routines so that the dollars truly follow the child, and in a timely way.

• Colyn Ritter, meanwhile, loves the new annual report about West Virginia’s Hope Scholarship program. He’s right that there’s lots to like, including demographic breakdowns about participants, detailed information about micro schoolers, and an in-depth look at spending patterns. (Amazon is the big winner so far.) But note the glaring absence of anything related to student learning. Is that too much to “Hope” for?

• Hope is on Mike McShane’s mind, too, in his case as applied to Catholic schooling in America. He reports on a new Apostolic Letter from Pope Leo that lays out his vision for Catholic schooling. Mike is hopeful because, after decades of decline for Catholic school enrollment, the burgeoning educational choice movement is leading Catholic schools to grow again. Mike writes:

  Using data from the National Catholic Educational Association and looking at the top ten states by Catholic school enrollment over the decade from 2013–14, [Ron Matus and Lauren May] found that Florida was the only state to see an increase in enrollment. While the total U.S. Catholic school population declined by 14.2 percent over that time period, Florida’s Catholic school enrollment grew 9.2 percent.

  Florida is not alone. Iowa Catholic schools saw an enrollment increase around 4 percent from 2023–24 to 2024–25.

  What do these two states have in common? If you’re reading this Substack, you’ve probably already figured it out: robust educational choice.

---

• After initial community resistance to Houston’s state takeover, a recent survey indicates families increasingly believe the district is heading in the right direction. —Naaz Modan, K–12 Dive

• “It’s not what they’re reading that matters; it’s that they’re reading.” So writes Erich May, a former English teacher and current superintendent in Pennsylvania who wants teachers to stop assigning “boring” books like those written by Shakespeare and Jane Austen. —Education Week

• “The small Minneapolis charter school where Renee Good sent her 6-year-old son has switched to online learning out of safety concerns after receiving multiple phone and email threats.” —Becky Z. Dernbach, Sahan Journal

• Education Week’s deep dive into school funding disruptions under Trump reveals widespread cuts related to stated goals like “shifting school funding responsibilities to states” and “cracking down on programs serving immigrant students.” But it also reveals reductions to initiatives the administration claims to support. —Mark Lieberman, Education Week

• “From Head Start to civil rights, 8 ways Trump reshaped education in just 1 year.”—Linda Jacobson et al., The 74

---