What’s the quality of education like at Brooklyn’s P.S. 307?
This was the elementary school where New York Times writer Nikole Hannah-Jones sent her daughter. In a much-discussed essay, she described low expectations, an indifferent administrator, and disappointing outcomes. But that was several years ago, and schools can change as staff turns over. Indeed there is a relatively new principal (who declined an interview request).
So imagine you’re the parent of a child zoned for P.S. 307 trying to figure out how the school is doing now. Or you’re a city official who wants to determine whether it needs urgent intervention.
You might start by looking at math and reading test scores.
In 2025, only a quarter or so of P.S. 307’s students were proficient on state exams, less than half the city average. This is alarming, of course. Yet raw scores like these are not good measures of how well schools are educating their students since there are so many other factors at play. The vast majority of P.S. 307 students come from low-income families and nearly half qualify for special education.
To account for this, many experts suggest looking at how much progress students make at a given school. Yet growth scores produced by New York city and the state tell different stories about P.S. 307. By the city’s measure, the school is well below average, but by the state’s it’s slightly above.
At this point if you’re the parent or the city official, you’re understandably confused. So what is going on here?
It turns out that measuring elementary school growth is quite difficult. This is a profound and under-appreciated problem for the parent trying to choose an elementary school and the policymaker trying to distinguish between good and struggling schools.
Over the years, researchers have developed complex statistical models to answer an essential question: How well do students perform in a given school relative to what would be expected in other schools? The difference between expectation and reality is a school’s “value-added” score: an estimate of the school’s contribution to student learning. These metrics have been much-debated, especially when applied to teachers, but there is evidence that they are accurate and meaningful in certain contexts.
A few years ago, informed by this research, New York City schools started to calculate and publish value-added or “impact” scores for most of its schools. These measures are relatively accurate in middle and high school, says Peter Hull, a Brown University professor who has been involved in a number of studies measuring the effectiveness of city schools. But he and colleagues have not studied the elementary model (results of which I featured in a recent newsletter).
Other researchers who reviewed New York City’s elementary impact measure raised questions about its validity. The key concern is that it does not use student scores in third grade as a baseline to measure progress in fourth and fifth grades. “You’ve got to look at the distance traveled,” says Dan Goldhaber, an education researcher and value-added expert at the University of Washington.
New York State’s metric, on the other hand, tracks year-over-year growth based on students’ prior-year test scores. Both Goldhaber and Tom Kane, a Harvard professor, say they would be inclined to put more trust in the state’s growth measure, which rates P.S. 307 as a bit above average.
Yet even the state’s approach has significant limitations. Most importantly, it only judges elementary schools by students’ progress in fourth and fifth grades — it ignores any success (or failure) before then.
Such a measure could penalize a school that is effective in early grades because it might make it harder to see growth in later years. If this value-added measure is used in high-stakes evaluations, there would be incentives to move the best teachers out of younger grades (which some research showed happened in the No Child Left Behind era).
City officials say it’s these concerns that led them to design the model as they have, accounting for demographics like poverty but not third grade scores. They note that students with disabilities and English language learners may have baseline testing data from separate exams.
Experts say there is no easy way to judge elementary school growth since testing starts in third grade. “It’s a hard problem,” says Hull of Brown. High school models also have limitations because testing is only mandated in a single grade.
In the end, there is no simple, quantitative answer about whether P.S. 307 — or any elementary school — is helping students make academic progress.
We want parents to have useful information for choosing schools, and we want policymakers to know which schools are thriving and which are wallowing. Yet despite the proliferation of data and increasingly sophisticated statistical methods, it remains difficult to untangle an elementary school’s effect from the many other factors that shape student learning.
Of course, parsing test scores is only one way to judge a school. Another is to talk to parents. I recently went to P.S. 307 to do so — I’ll share what I learned next week.
Reach me at [email protected]. Forwarded this email? Sign up here to receive the Chalkbeat Ideas newsletter.
Looking for your next read? Check out these other great newsletters.
