If you could find 25 hours of professional development time for every teacher in your school, what use of those hours would have the biggest impact on student achievement? That is the question Professor Dylan Wiliam puts to school leaders, and in this Community of Practice session for the Department for Education, South Australia, he sets out why the answer is classroom formative assessment.
Recorded live on 9 September 2026, the session was hosted by Pádraic Hogan of Nurture with educators and leaders from the Mount Gambier 2 Portfolio, whose schools have spent the past six to nine months shaping Nurture Signal as a design partner. Dylan gives a mini keynote on the design and evidence behind the Embedding Formative Assessment programme, then takes questions from the group on data, feedback, parents, and the hardest problem of all — finding time.
Who it is for
This session is designed for educators and leaders in South Australian public schools, including:
- Classroom teachers in primary and secondary settings
- Principals and site leaders
- Education directors and portfolio leaders
- Teaching and learning, curriculum and assessment leads
- Anyone planning a formative assessment programme across more than one site
What the webinar covers
Why the programme is designed the way it is
Dylan opens with 20 years of work on effective, scalable teacher professional development. Research showing that something improves achievement is beside the point, he argues, unless we also know whether it can be done in real classrooms, how much teacher time it takes, and how much extra learning it buys. The way to improve a system is not to replace teachers with better ones, but to invest in the teachers already in the building.
Expertise as habit, not knowledge
Much of what expert teachers know is implicit — closer to knowing how to ride a bicycle than knowing how to solve an equation. Some things can be made explicit and simply told to teachers: a prime number is a number with exactly two factors, which settles what the usual definition leaves open. But a great deal of teacher expertise is explicit and still hard to enact. Every teacher knows they do not wait long enough after asking a question; teachers guess the average wait time is about three seconds, and Mary Budd Rowe’s research put it at 0.9. Reminding teachers of that research, Dylan says, has about as much impact as reminding smokers about the harm of smoking. It is not a knowledge problem, it is a habit change problem — what Mary Kennedy calls the knowing-doing gap.
The design principles
- In-principle scalability — nothing built in that would make it impossible to run across tens of thousands of classrooms
- A single model for the whole school, so students meet a coherent approach rather than one reform in maths and another in science
- Formative assessment as both generic and subject specific — what makes a good science question differs from a good history question, but getting a response from every student rather than the confident volunteers is a generic idea
- A real account of scale — depth into practice, sustainability, spread across the school, and a shift in reform ownership so teachers eventually set their own agenda
- Content before process — coaching triads and critical friendship groups are fine processes, but nothing in them guarantees teachers spend the time on something that works
The evidence
The Embedding Formative Assessment programme — whole-school, teacher-led, two years, five strategies, detailed resource packs, monthly meetings and peer observation — was evaluated in a randomised controlled trial funded by the Education Endowment Foundation and run by the National Institute of Economic and Social Research. Dylan walks through why the design matters: pre-registration to prevent outcome switching, randomisation, and an intention-to-treat analysis rather than cherry-picking the schools that implemented it well.
The researcher’s question is, does this programme work if it’s done properly? The school leader’s question is, if we adopt this programme, will it work?
The result: students in the formative assessment schools made 25% more progress, achieved by changing how teachers spent about 1% of their time, at a cost of roughly A$2 per student per year. For comparison, class size reduction buys around 12% more learning at something closer to $5,000 per student per year.
In the discussion
- Data: moving from data-driven decision making to decision-driven data collection — start with the decisions you need to take, then work out what data would help you take them better. Be inclusive about what counts as data, and remember that the easiest person to fool is yourself.
- Feedback: in around 38% of the studies Dylan cites, feedback actually made learning worse. The only good feedback is feedback that gets used by students; good feedback is whatever gets you more of what you want from your learners.
- Contingent teaching: the shift from hearing back from the usual suspects to hearing from every student — “you matter, you belong, I need to hear from you so I can make decisions that reflect your learning needs”.
- Student agency: Dylan’s goal has changed. Higher achievement, but not at the cost of agency — students should leave at 18 with the curiosity they arrived with at five. Building agency, he argues, is now the major purpose of formative assessment.
- Parents: why the second Embedding Formative Assessment book includes a sample letter home, and the framing that works — the exam is the Olympics, and this is the training. Feedback opens a three-cornered conversation between student, teacher and parent.
- Finding time: leaders fail to clear space because they look for bad things to stop doing, and there aren’t any. The essence of effective leadership is stopping teachers doing good things so they have time to do even better things. Dylan points to Making Room for Impact, written with Arran Hamilton and John Hattie, as a discovery process for finding that space — hard, because we are cognitively primed for addition, and because teaching is an infinite job.
Designing better checking for understanding
Dylan closes on the two steps that make a check for understanding worth running:
- Students with the right thinking and students with the wrong thinking must not give the same answer. If they do, it isn’t a good question.
- Refine the wrong answers so different misconceptions produce different responses.
Do both, and something changes in the teacher. Instead of listening evaluatively — scanning for the right answer, where every wrong answer is equivalent and gets a “nearly, close, nice try” — teachers start listening interpretively, reading which answer a student chose as a clue to the difficulty they are having.
Where this fits with Nurture
Pádraic Hogan sets out the two things Nurture has built alongside this work:
- Nurture Learn — a self-paced professional development programme developed with Dylan’s guidance, covering the five key strategies of formative assessment with techniques teachers can apply in their own classrooms, with departmental resource packs to follow
- Nurture Signal — engagement check-ins built on Dr Amy Berry’s Continuum of Engagement, disposition check-ins, and checking for understanding every few minutes rather than every few weeks, including hinge questions that capture each student’s reasoning and a banded read on whether to move on, recap or reteach
The session also looks ahead to 2027: growing from a proof of concept with a handful of Mount Gambier 2 Portfolio schools into a larger, more comprehensive pilot, with a dedicated teaching and learning lead in South Australia, protected time for professional learning, and an evaluation framework designed in from the start.
Speakers
Professor Dylan Wiliam
Emeritus Professor of Educational Assessment at University College London and Strategic Advisor to Nurture, Dylan Wiliam is one of the world’s leading authorities on formative assessment. His research focuses on supporting teachers to develop their use of assessment in support of learning.
Pádraic Hogan
Co-Founder and COO of Nurture. Pádraic hosts the session and shares the story of Nurture’s work with the Mount Gambier 2 Portfolio, from proof of concept to the 2027 programme.
Key takeaways
- Why classroom formative assessment is the most cost-effective use of professional learning time we know of
- What the Education Endowment Foundation trial actually measured, and why intention-to-treat analysis matters when judging any programme — including AI products
- Why professional learning has to target habit change, not knowledge gaps
- How to design hinge questions that tell you what students are thinking, not just whether they are right
- Practical guidance on protecting teacher time, involving parents, and building student agency
On-demand webinar
Watch the full recording
Get free access to watch this webinar now.
You have free access to this recording.