What Works (and Doesn’t) in Education?

I had breakfast with John Hattie and Peter Hill yesterday. Along with another thirty souls. We talked about what works and what doesn’t in education. I’ve written out the best of it below.


Scale Change

  • Noted that all the evidence of meta-analyses shows that all of the most important factors in kids outcomes are teacher-related, yet almost all policies and systemic efforts target schools and regulatory changes. System efforts right now are almost always misguided. John Hattie is now getting obsessed with the idea: how do you scale up great ideas and great practice?


  • The most crucial thing for teachers and educators to talk about and align on is impact. This is the most important and should be the most frequent conversation in education. What is the impact you want in your class or your school? How will you know that you’re seeing the impact that you want? How will you achieve that impact? This is opposed to current conversations which tend to be about content.


  • We obsess about school choice for kids at the systemic level. Yet school effect on kids’ outcomes is only 7% when controlled for teachers, and teacher effect is 40%. We should really be obsessed with the choice of which classroom kids are in within a school. Anyway, the key point is – and we already know it – individual and aggregate teacher quality is the vital component of great education.


  • We think about this incorrectly. It’s a mistake to get kids to self-asses, the data shows us that they already know what level they are at. Instead we should think of assessments as tools for teachers to gain insights on how well they (the teachers) are doing. The assessment shouldn’t be to evaluate the kid, but to evaluate the teacher. This then shows the teacher how they need to personalise things for different kids. There’ll be good data tools to help teachers automate some of this soon.


  • Tools are being developed to help the teacher leverage ‘Visible Learning’. If data systems can track what kids are doing, then the teacher can SEE what is going on for all kids in the class. Currently 80% of what happens in class isn’t noticed by anyone. That’s a lot of lost learning.
  • This will change the teacher role will change significantly. Intelligent and rapid data analysis by machines will facilitate personalised read outs about learner needs very quickly. The teacher will therefore become increasingly engaged in the personalisation of learning. Assessment will generally be a formative rather than summative tool.


  • It’s incredible that failure in the classroom is perceived in the way that it is now. Imagine how few classes there are where failure and ‘not knowing’ is privileged. And yet we can only truly learn when we’re failing and exploring things we don’t know. Current common idea of a good learner is a kid who learns quickly without trying. Yet this is that absolute antithesis of what we think of when we consider what a good learner is. We must absolutely privilege what we don’t know… That space is the best prompt for learning.


  • It’s not how you give it, it’s how you receive it. As a giver of feedback, how can you ensure that it is received well by the intended recipient? Getting kids or teachers to receive it well is the key and this requires ‘where-to-next?’ feedback, rather than any other kind. People can’t act on much feedback (studies suggest a limit of 3 seconds worth per day of actionable feedback that may change someone’s behaviour).
This entry was posted in EDUCATION. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *