There are studies you read and immediately think: this is going to be interpreted in ways the authors did not intend. This is one of them. A new paper in the British Educational Research Journal analyses more than two decades of PISA data (2000–2022) across eleven European countries and asks a seemingly simple question: do students in publicly funded private schools perform better than those in public schools?

The short answer from Priya Maurya and colleagues is: sometimes. A part of the much longer answer is: it depends. And that longer answer is where things get interesting.

If you look only at the raw scores, students in publicly funded private schools often outperform their peers in public schools. That pattern appears in several countries. But the authors are careful not to overstate what this means. They explicitly talk about associations, not causal effects. As readers of this blog will know, PISA does not allow for strong causal claims. Students are not randomly assigned to schools. Families choose, schools select, and those processes are far from neutral.

You can see that in the data. Students in publicly funded private schools tend, on average, to come from more advantaged backgrounds. Once you take those differences into account, the performance gaps often shrink. In some cases, they disappear. In other words, part of the “advantage” is not about what schools do, but about who attends them.

Another layer we often tend to lose in the debate. The pattern varies substantially across countries. In some systems, there are modest positive differences. In others, there is little to no difference. And specifically in  Luxembourg, the pattern even goes in the opposite direction. This is not a minor detail. It suggests that the issue is not “public versus private” as such, but how education systems are organised in practice.

That is also how the authors frame their findings. Differences between school sectors follow institutional configurations rather than simple ownership categories. Governance, regulation, funding arrangements, and degrees of autonomy all matter. Treating publicly funded private schools as a single, uniform category across countries obscures more than it reveals.

Socio-economic background adds another important nuance. The study finds that positive associations are often stronger among students from middle- and higher-socioeconomic backgrounds, while effects among more disadvantaged students are smaller or less consistent. This is not definitive evidence of mechanisms of inequality at work, but it is a pattern that raises important questions about access and distribution.

All of this helps explain why studies like this tend to become contentious. It is easy to read them as confirmation of prior beliefs. Some will see evidence that private provision “works better”. Others will interpret the same results as reflecting selection and segregation. Both readings draw on elements in the data, but neither captures the full picture.

What the study actually does is shift the question. Not: Which type of school is better? But: under what conditions does a particular type of school seem to work better, and for whom? That is a less comfortable question, because it does not lead to a simple answer. It forces you to consider how different elements of an education system interact, including student composition, governance structures, and policy design.

There is one final nuance worth noting. Looking at the period around the COVID-19 pandemic, the study does not find a consistent advantage for one type of school over another. In some contexts, publicly funded private schools appear to have done better. In others, public schools did. Again, the pattern is mixed. And again, context seems to matter more than labels.

Perhaps that is the main takeaway. Education systems rarely behave in the clean, predictable ways we might prefer in public debate. There is no single lever that reliably produces better outcomes. Not public versus private, not autonomy versus control, not choice versus equity. What matters is how these elements are combined, and in which context.

That may be less appealing than a clear headline. But it is closer to what the evidence actually shows.