I first came across this meme in a talk by Dylan Wiliam, but here is a quote from Andreas Schleicher, saying something similar:
But just about a quarter of student performance variation lies between schools in the UK, while three quarters lie within schools, a share that is much larger than in other countries. In other words, educational underperformance isn’t just about poor schools in poor neighbourhoods, but about many schools in many neighbourhoods.
Wiliam uses this as an argument that what goes on in any particular classroom matters [most]. Teachers matter. My initial reaction was to imagine that such effects will be much smaller in highed, but I can conceive of ways in which I might be wrong. If you think about medical education, there is (inevitably) focus on macro factors; course design, how the course is divvied up between specialty groups, IT support, and so on. But although lip service is paid to promoting ‘medical teachers’ (most of which is, in my opinion, misguided), we know little about what actually goes on in our classrooms, and even less about any relation between inputs and outputs. I am not certain we will ever get the appropriate figures, but think we should act as though we know the answer.