The urgency of how: Let’s design our systems to learn

“It is amazing that the systems that we entrust with managing our own children’s learning are terrible at learning themselves.” 

So Harvard University economist Tom Kane bluntly and accurately framed a core impediment to meaningful advances in student learning across the country. He recently made this statement to a roomful of the nation’s top tutoring providers and researchers as they took in the urgency laid bare in the results of Kane’s latest research with Stanford Professor Sean Reardon on K-12 student learning during and since the pandemic. 

While Kane and Reardon found that student achievement did improve between Spring 2022 and Spring 2023, students are still well below where they were in 2019. They estimate it will take at least another year of recovery in math and two more in reading to get back to pre-pandemic student achievement levels. Alabama is the only state where average student achievement in math exceeds pre-pandemic levels; only Louisiana, Illinois, and Mississippi have above pre-pandemic average achievement levels in reading. 

And a closer look at these and all states shows yawning achievement gaps along lines of poverty and race. 

Layer that with the fact that school districts face a looming funding cliff: They have until the end of September to determine how they will spend their remaining federal relief funds, or they will lose them. As of January, districts had yet to spend about $51 billion of the initial $190 billion, according to Kane’s research. President Biden proposed a slight bump in education funding in his budget, but it’s unclear how Congress will act.   

What can we do? the tutoring providers asked Kane. 

Central to this quandary, Kane noted, is that many educators still don’t know which interventions are working for which students and why. Many states and districts have failed to analyze which programs helped students the most.

To accelerate and make sustained progress and ensure we are spending limited funds efficiently and effectively, we need to redesign systems to learn. 

That means systems  – whether states, districts, schools or classrooms – reorganize themselves to collect relevant information to understand how and whether strategies and practices are working and then use what they learn to adapt their practices and policies to better meet the needs of every student.

That means building the capacity of leaders to enable this continuous learning, and to adjust and adapt priorities and strategies over time. 

That means seeing interventions as a helpful driver, but not THE driver, and instead shifting focus to creating a system that can make many interventions successful in different contexts. 

Had we been doing this the last few years (or decades, for that matter), we would undoubtedly have made better decisions about how to deploy increasingly scarce resources. 

We know from Kane and Reardon’s report that some things worked over the last year: They found that students in all but one state (Oregon) improved in math achievement; students in Tennessee, Pennsylvania, and Mississippi improved in math by more than a third of a grade level; and students in Mississippi and Illinois made up more than a third of a grade level in reading. 

Sustaining and spreading these improvements requires learning how they were achieved, putting processes in place for effectively assessing implementation and making changes to suit the students in front of them. Hard work, indeed, but these are the steps that systems that learn take to avoid lurching from one intervention to the next in the hope of making widespread improvements.

Together, let's begin to focus more of our energy on our public schools and districts and state agencies' capacity to deliver any intervention even as we remain focused on the quality of interventions themselves. Through this approach of supporting system learning we will see widespread and sustained improvement to student learning.