Keynote I
7 Oct, 09:30 — Chair of Learning Sciences, Marlowe University London
The marking machine: what automated assessment learned from teachers, and what it forgot
Drawing on a four-year study across 62 secondary schools, Nwosu shows where machine marking outperforms panels of teachers, where it reliably fails, and why the failures cluster on the students who can least afford them.
Keynote II
8 Oct, 09:30 — Senior Research Fellow, Norwegian Institute for Educational Research
Twelve years of one-to-one devices: the Nordic evidence, reconsidered
Sandvik revisits the longitudinal data behind the Nordic reversals, separating the effects of the devices from the policies around them — and argues the strongest finding is about what schools stopped doing.
Keynote III
8 Oct, 15:45 — Professor of Curriculum Studies, National Institute of Education, Singapore
Small data, close reading: classroom research after the analytics boom
Tan makes the case for a return to close observation — video, transcripts, teacher diaries — presenting three studies in which the interesting result was invisible to every metric the platform collected.