Let me start this post with a confession. I’ve had a somewhat complicated relationship with Gert Biesta’s work ever since I first heard him speak live, back when I was still working at Artevelde University of Applied Sciences. Often when I read something he wrote, I will think: yes, of course. But I often notice that I am much less convinced by the position he argues against. A long post as reaction, discussing also evidence-informed education.

I had that feeling again with his new article, What kind of research for what kind of education? Lessons for the OECD and others. In it, Biesta responds to a 2025 OECD report on how research institutions and initial teacher education institutions can support the use of research in education.

I agree with many of Biesta’s points. For example, he argues that it is too limited to think about the relationship between research and education only in terms of knowledge about what you can do to achieve particular outcomes. He distinguishes between instrumental knowledge and orientational knowledge. Instrumental knowledge concerns possible relationships between our actions and their consequences. Orientational knowledge, on the other hand, helps teachers to orient themselves in their practice by answering questions such as how do we look at students, schools, inequality or power? And, also, what should education actually be about? I do think this distinction can be useful.

The same applies to his warning that not all educational research needs to be close to daily practice. This may sound bizarre to some, but I agree. In his model, Biesta combines these two types of knowledge with another distinction: research close to concrete problems in practice versus research that deliberately keeps some distance.

These distinctions create a quadrant (this sounds almost like someone in HR) with four forms of educational research. According to Biesta, all four have a legitimate place. Again, I can go a long way with him here. Fundamental research does not need to produce a useful tip for a teacher next month to be valuable. Some of those insights are much needed for the rest of the quadrant. So far, so good.

Is this really what the OECD says?

But this is where things start to bother me. Biesta argues that the OECD report mainly treats the relationship between research and practice as some kind of logistical problem. The knowledge is there, but practice does not sufficiently reach it. Hence the term knowledge mobilisation: we need to ensure knowledge moves from research into education. According to Biesta, this metaphor suggests that the main task is to improve the “flow” of already available research knowledge towards policy and practice.

But the OECD report itself uses a considerably broader definition. In the report, knowledge mobilisation involves more than making research accessible. The OECD explicitly mentions generating research relevant to policy and practice, building relationships among research, policy, and practice, and increasing people’s capacity to engage with research. It is explicitly a two-way process, or perhaps more accurately, a multi-way process. This becomes clear elsewhere in the report, which describes how connections between teacher education institutions and schools can help identify research needs emerging from practice.

You have to admit, that is rather different from the reasoning that “researchers have produced knowledge and now we just need to get it to teachers.”

This makes clear what I already mentioned in my introduction. Biesta often offers an interesting alternative. But the question is whether the opponent he positions against would recognise themselves in the position he attributes to them.

Causality does not mean certainty

This mechanism becomes even clearer when Biesta discusses effectiveness and causality. He first makes an important point. Educational research can never guarantee that a particular action in a particular situation will produce a predetermined result. What worked elsewhere does not necessarily work in exactly the same way here. A teacher therefore needs to consider the specific students and situation and use professional judgement to decide what matters. Biesta therefore describes instrumental knowledge as knowledge about possibilities rather than certainties. It is difficult to disagree with this, although I do think context shouldn’t lead to complete relativism.

But then Biesta goes considerably further. He argues that in the domain of what varies, such causality does not exist, and that the suggestion that research can tell us “what works” therefore rests on a fundamental misunderstanding of education. This is where he loses me. Who says that causality, by definition, has to be a synonym for certainty?

When a randomised experiment shows an intervention has an average effect, this does not mean every student will experience the same effect. There is a reason why we talk about “evidence”, not “proof”. Effects can differ between students, teachers, schools and contexts. That’s also why we are talking about an average in the first place. We can be uncertain about an effect’s size, and something may work better in one context than another. This complicates causal claims, and it’s one reason I keep discussing the nuances of studies, including RCTs, in almost every post on this blog. But more complicated and less certain does not mean impossible or unimportant.

In his work, there seems to be a shift from “research cannot provide certainty about what will happen to this student in this situation” to “research cannot tell us what could work”. But logically, the second does not necessarily follow from the first.

This argument is not new in Biesta’s work. In his 2022 book on educational research, he also described education as an open system in which teachers’ and students’ agency makes a simple causal approach problematic. Experiments can, still according to Biesta, tell us what has worked in the past, but we cannot simply infer from that what will work in a new situation. Again, I agree with the latter because of the word ‘simply’. But it does not follow that research cannot teach us anything about causal effects. Having no certainty about an individual future situation is not the same as having no knowledge about the probability of particular effects under particular conditions. Btw, if this was the case, it would be impossible to teach if you take this line of thinking even further.

Indeed, the question Biesta himself raises, It worked there. Will it work here?, is not an argument against evidence-informed education for me. Quite simply because it is almost at the heart of it.

Research does not decide what good education is

I keep alternating between agreeing and disagreeing with Biesta in this post, so here is another point where I think he is right. Effectiveness is never the only question we should ask in education. He gives the example of punishments and rewards. You can use them to achieve particular behaviours or outcomes, but the way you do so also communicates something. Therefore, asking if something is effective is not enough. It is key to also examine whether it is ethically and educationally desirable.

If research shows that approach A is more effective than approach B for goal X, this does not mean that we should automatically use approach A. We still need to decide other things too:

• whether X is a goal we consider important,
• what other consequences A may have (I like the idea of possible side-effects), and
• whether we consider this way of acting educationally appropriate.

But again: is this a problem for evidence-informed education, or is it actually an essential part of it? In my view, it is only a problem if you define evidence-informed practice as research determining what teachers should do. And I do not recognise that definition.

Perhaps we agree more than Biesta thinks

As you can see, I often agree with Gert. Let me summarise some of the points that regular readers of this blog will recognise:

• Research does not give teachers recipes.
• Context matters.
• Professional judgement remains necessary.
• Research can produce different kinds of knowledge.
• Effectiveness is not the same as desirability.
• You cannot simply derive the aims of education from empirical research.

Perhaps the idea that not all valuable research needs to be immediately applicable has received less attention on this blog lately, but believe me, I agree with that too. What I do not see is where the contradiction with evidence-informed education lies. Biesta has often been critical of evidence-based education. But much of what he defends here is precisely what I recognise in my own understanding of evidence-informed education. Aren’t these precisely some of the reasons why the shift from evidence-based to evidence-informed took place?

Research is one source for professional decisions, alongside practical knowledge in the broadest sense. It can tell us about the possible consequences of our choices, reveal patterns that are hard to spot in everyday practice, or show that some of our intuitions may be wrong, as our work on educational myths has often shown. Other research may simply help us look at our practice differently. Teachers then still have to make sense of all this in their own context, with their own expertise, aims and values. Research doesn’t replace professional judgement. Hopefully, it improves that judgement.

To be clear: in this article, Biesta is not directly targeting evidence-informed education, but the OECD report and a particular conception of “what works”. But who actually still defends the position being argued against here?

Perhaps some people still think that research can provide certainty or that causal effects should be the same for everyone. And I know people outside education who hope context doesn’t matter, though even they usually know better. If you think teachers should simply implement research findings, you are probably not an education person.

Perhaps such views exist. But I rarely encounter them among serious advocates of evidence-informed education.

And so I am left with the question: who is Biesta actually arguing with?