Statistical grade boundary setting approaches

Last updated:

Alpha Plus Consultancy Ltd report for the IB

Every exam session, the IB carries out a large number of grade award processes to determine the final grade boundaries for each subject. Grade award is a rigorous process that involves the analysis of a range of different evidence to ensure the fairest and most comparable outcomes for students. This evidence includes Statistically Recommended Boundaries (SRBs) which provide a statistical estimate of where the grade boundaries should be to maintain standards from one year to the next. Whilst SRBs offer useful statistical guidance in many circumstances, there are also certain situations where they work less well, such as when there are very few candidates in a subject or when introducing a new curriculum or assessment model. In this study, AlphaPlus reviewed different statistical methods for calculating recommended grade boundaries and considered their suitability for different scenarios in Diploma Programme and Middle Years Programme assessments. It concluded that while the current approach works well in many cases, there are some grade award contexts where alternative methods might be more suitable.

Executive summary (PDF, 591 KB)

Full literature review report (PDF, 1.4 MB)

Full simulation report (PDF, 2.4 MB)