Read on Twitter
UCAS predicted grades deserve more attention in resolving #alevel2020 problems in university entry. Correctly handled, they give accurate and testable estimates of what students with a firm offer would likely have got relative to other offer holders in a normal exam year (1/4)
The models work well because the information encoded in the grade profile / type of course offer - not available to Ofqual - is important to how people do. Background such as school, sex or area are not used, but estimates are unbiased by these. Works for BTEC/A level too. (2/4)
We outlined this broad approach in March. https://wonkhe.com/blogs/we-can-make-admissions-work-without-a-levels/ . It would have reduced the uncertainty in university entry resulting from the different task Ofqual had. And, with some relaxing of number controls, would have let universities resolve problems. (3/4)
Applicants who fear they have lost places would perhaps be better served by a component of the appeal process that draws on the UCAS predicted grade *set* and university offer to spot those who had an e.g. 50%+ chance of meeting conditions, not an atomised grade approach. (4/4)
