“It is a wretched taste to be gratified with mediocrity when the excellent lies before us.” Disraeli

What could be wrong with wanting students to master difficult content? Nothing. For the most part, the aims of mastery curricula are admirable. Ensuring all students have fully grasped conceptually difficult content is a hard but worthy aspiration. My problem is that, in practice, mastery values the here and now over the future, and in so doing may be in danger of short circuiting the outcomes it seeks to embed.
The research conducted so far shows some promise. The EEF Toolkit report concludes that mastery learning offers, “Moderate impact for very low cost, based on moderate evidence.” But there seems to be little agreement on what mastery actually means. Amidst all the competing definitions, this one from the Assessment Without Levels commission has gained some traction:

…in mastery learning, ‘learning is broken down into discrete units and presented in logical order. Pupils are required to demonstrate mastery of the learning from each unit before being allowed to move on to the next, with the assumption that all pupils will achieve this level of mastery if they are appropriately supported.

Here’s why I think it’s potentially flawed.
Learning is the retention and transfer of knowledge and skills. If students can’t remember what they’ve been taught elsewhere and later, then they can hardly be said to have learned, let alone mastered it. So how can we know whether students have learned something? We can’t. All we can ever do is infer from their current performance what they are likely to be able to do elsewhere and later; we cannot see into the future. Of course, some performances provide more valuable evidence than others. A test six months after instruction allows us to make a much better inference than answers to questions in the lesson where material is first taught.
The biggest problem with some incarnations of mastery curricula is that once students are judged to have ‘mastered’ a body of knowledge, teachers then move on to more challenging content. But as we learn so too do we forget. Although the rate at which we forget is highly variable, the fact that, over time, we will be unable to retrieve the majority of what we once knew is one of the most robust findings from decades of research. Fortunately, we can take steps to reduce and possibly even arrest this decay but only if we assume that what students appear to have mastered will likely be forgotten.
The troubling and counter-intuitive finding is that measure that increase current performance often seems to undermine future learning. If we seek to improve students’ performance in the here and now we run the risk of producing in them the illusion of knowledge. When material is familiar we believe we know it when actually all we really know is that we once knew it. The memory of mastery endures while the substance of what we think we know falls away without us noticing. I remember going to see a lecture by the poster boy of British physics, professor Brian Cox. He gave a fascinating talk on the quantum universe and I left buzzing. A few days later I was telling a friend about how much I learned. “So, tell me,” my friend asked, “what have you learned?” I thought for a few puzzled moments before admitted that all I could remember was that the talk had been really good and that I’d enjoyed it.
This matters because when we think we know something we stop thinking about it. When we know we don’t know something we carry on thinking. This is often unbidden; we sleep on it, we put it on the back burner, we chew it over, and sometimes, in what can seem like a sudden epiphany, but is actually the result of slow, painful attempts to integrate new information, we ‘get’ it. The trick, if there is a trick, is to expose students to the extent of their ignorance so they continue thinking about what they don’t know.
At any one time, students think and believe more than one thing. These ideas compete with the older, more established idea usually winning out over the new idea. When we teach, students’ performance improves; the new idea comes to the fore. They seem like they’ve mastered it, but this is, as often as not, mere mimicry. Conceptual change, the pressure to transform or revise misconceptions as new understandings are learned is tricky. The old idea re-emerges and the new fades into the background. When we refuse to take students’ current performance as evidence of mastery, we acknowledge the difficulty of battling misconceptions. Change is gradual. It takes time and continual reintroductions of new ways of thinking for the change to stick and become permanent. My fear is that many mastery programmes miss this important truth and run the risk of making false and potentially damaging assumptions.
Here, Jo Facer sets out her vision of what mastery learning should look like. There’s lots to like and plenty I agree with. She acknowledges the need for quizzing, overlearning and distributed practice to build long-term memory and rightly observes that memory is complex. But there are a few aspects I’d question. The first is that I’m fairly sure that Jo’s definition of a mastery curriculum will not be shared by many schools who feel they offering something with the same name but substantially different in terms of delivery. But that’s an aside. I have some concern over statements like these:

We need to not ignore the possibility that students have not understood what they have read: teachers need to make use of comprehension questions to ensure students are showing they understand

There’s nothing wrong with using comprehension questions to find out students’ current levels of understanding, but it would be a mistake to think this “ensures” anything. Many people argue that it’s better to know that not know what students are thinking and, if it were possible to really know what students thought, I’d agree. Correct answers to comprehension questions fool us into assuming mimicry is evidence of learning.
But much more problematic to my mind is the idea of a 5 part mastery lesson. Jo writes that a mastery lesson should

…begin by recapping on students’ prior knowledge, before reading or instructing pupils (ideally both) in the new knowledge. Questions should be used to check for understanding as well as to stretch and challenge. Students then need a period of time for deliberate practice. Finally, homework should support the core purpose of the mastery curriculum: committing the most important knowledge to long-term memory.

Every lesson? I’ve written before about the problems with planning lessons and think this structure, while appropriate sometimes, should be resisted as a template for every lesson.
In summary, ‘mastery’ runs the risk of becoming a weasel word, meaning something different to everyone who uses it. The idea that teachers move on when children master content is certainly preferable to teachers moving on without children mastering content, but it may not get the durable, flexible results we really want.