12 July 2021

Find out about the key findings from the OSR review of statistical modelling following the 2020 awarding of exam grades.

In 2020, the awarding of exam grades in the UK brought the use of algorithms and statistical modelling to the fore of public consciousness in a way that had not been seen before, despite their rise in use over many years. It raised questions about the public acceptability of statistical modelling and communicating with the public on these subjects. The Office for Statistics Regulation (OSR), the regulatory arm of the UK Statistics Authority, is the body responsible for upholding public confidence in statistics. In the months that followed the awarding of grades, OSR undertook a review of lessons learned from the approach to developing the models that decided exam results. Using exam grades as a test case, OSR sought to understand – in a world of abundant data, what can be learned about the real-life consequences of using algorithms?

Watch an online lecture delivered by Head and Director General for Regulation at the Office for Statistics Regulation, Ed Humpherson. Organised by the University of Bath Institute for Policy Research (IPR) and UKRI CDT in Accountable, Responsible and Transparent AI. In this lecture, Ed Humpherson discusses the review’s findings, what Government and public bodies should learn from the exams case, and what it means for the use of algorithms and statistical models in public policy going forward.

Watch the lecture about the review's findings on YouTube