Editor’s note: This was first published in SCHOOLED, Mike Petrilli’s Substack, which hosts weekly education-reform conversations around K–12’s biggest topics.
Let me add some data from outside the U.S. that might sharpen the question Mike Petrilli posed recently on his SCHOOLED Substack: “Should tutoring play a big role in America’s schools going forward?”
I’ve lived in the Netherlands for decades, where I got my Ph.D. and hold an associate professor position in sociology at the University of Amsterdam. Perhaps relevant to this crowd: The Dutch welfare state may reduce the share of kids growing up in poverty—we can debate that. But within the remaining pockets of poverty, the welfare state clearly is not, by itself, creating conditions where truly disadvantaged kids learn. High-poverty schools here are less violent, fewer gangs, yes. But they remain lousy. As in the U.S., we’re seeing, for example, huge percentages of kids 4 years behind by age 12.
After following Match from the blueprint stage in the late 90s—and high-dosage tutoring (HDT) at Match from the early 2000s—I started advocating for bringing HDT to high-poverty primary and secondary schools in the Netherlands in 2008. I finally got philanthropic support in 2013, which led to the first HDT intervention in the country, at the primary level—as far as I know, the first in Europe.
The results across four randomized controlled trials (RCTs):
- First primary-level RCT: 0.28 (p < 0.01) standard deviation (SD). Implementation was mediocre; I couldn’t hire or fire and kept getting pushback.
- Started a non-profit dedicated to HDT. First secondary-level RCT: 0.44 (p < 0.01) to 0.72 (p < 0.01) SD depending on the test.
- Primary school in the poorest hood in Amsterdam, RCT: 0.3–0.65 SD depending on the model. (Dutch only.)
- Most recent secondary-level RCT: 0.19 SD. (Dutch only.) My sense is that, partly due to research design demands, we were forced to allow in too many students who were already fairly advanced given the curricular materials. But it could also be less fire in the belly, less talent. Hard to tell.
So what does this tell us about Mike Petrilli’s original question?
Liz Cohen’s ideas are compelling; I will read the book! As the essay Mike Goldstein and I wrote indicated, the lesson from the Netherlands is the same lesson you find everywhere: Tutoring can be transformative if you have enough hard-working, talented people running the intervention—and, I’d now add, if you have adequate fit between (1) what you offer (curriculum and tutor skills), (2) the skill level and needs of the kids, and (3) the research design/high-stakes standardized test you’re using to measure growth.
As I said on Mike Petrilli’s podcast, I think a lot boils down to this: When you’re working in overwhelmed schools, no matter what the program, a “tsunami of curveballs” is going to be thrown at you. So...do you have managers who will troubleshoot, again and again? I mean are they warriors? Or will they do what countless managers and teachers do in high-poverty schools all over the world—lower expectations, make excuses?
Here’s my harder conclusion: I didn’t get into this to achieve less than, say, 0.6 SD. We all know about fade-out versus persistence of treatment effects. Anything less than such a meaty effect size isn’t going to alter trajectories or increase objective life chances. Given all the political, organizational, and financial constraints, and the ongoing challenges to operational excellence on the ground, I’ve concluded it’s going to be damn hard—even on a small scale—to achieve and maintain tutoring interventions that generate anything like 0.6 SD, even when you can build coalitions willing to pay for in-person, relatively intense models costing around $5,000 per kid, per year. And I think you can imagine how difficult this is.
Scaling up, cutting costs, moving online to reduce expenses and solve sourcing problems... I think you need truly exceptional managers who can build and maintain an extremely tight culture.
Which brings me back to the “degree of difficulty” framing. The difficulty isn’t primarily in the intervention design. It’s in building and sustaining the organizational conditions where the intervention can actually rock. That’s a different kind of problem—and one that policy levers are poorly suited to solve.
To answer Mike Petrilli’s original question: Yes, tutoring can be “the next big thing”—but only school by school, crew by crew. If you’re warriors and want to try it for a few years, go for it. Just know what you’re signing up for.
