Ofqual algorithm skews student grade predictions

An algorithm developed by a UK government agency to predict student exam grades during the COVID-19 pandemic resulted in inaccurate and unfair results, and exacerbated socio-economic inequality.

In 2020, qualifications, exams, and tests regulator Ofqual devised a grades standardisation algorithm to address grade inflation and moderate teacher-predicted grades for A-level and GCSE qualifications in England. 

However, the algorithm faced significant criticism for a flawed design process, leading to inaccurate results; and inadequate accountability. It also resulted in multiple negative impacts, including: 

• Downgraded results. Nearly 40 percent of students received exam scores that were lower than their teachers’ predictions, jeopardising their chances of securing university spots.

• Disproportionate impact. The system disproportionately affected students from working-class and disadvantaged communities, who faced more severe downgrades compared to others.

• Private school advantage. The algorithm was found to have inflated the scores of students from private schools, potentially giving them an unfair advantage, exacerbating existing socio-economic disparities.

An Equality and Human Rights Commission report stated that the Ofqual’s algorithm may have a lasting effect on young people from certain ethnic minority backgrounds, disabled pupils, and those with special educational needs who were already disproportionately disadvantaged.

The incident resulted in the resignation of Ofqual's chief regulator. It also led the opposition Labour Party to accuse the agency of violating the UK Apprenticeships, Skills, Children and Learning Act 2009.

The UK government later abandoned the algorithm in favour of teacher-led assessments.

Regulation ⚖️

• Apprenticeships, Skills, Children and Learning Act 2009

Research, advocacy 🧮

• Mallett B. (2023). Reviewing the impact of OFQUAL’s assessment ‘algorithm’ on racial inequalities 

• Office for Statistics Regulation (2021). Ensuring statistical models command public confidence. Learning lessons from the approach to developing models for awarding grades in the UK in 2020

• Centre for Progressive Policy (2021). Is the algorithm working for us? (pdf)

Investigations, assessments, audits 🧐

• House of Commons Education Committee (2020). Getting the grades they’ve earned

News, commentary, analysis 🗞️

• https://www.bbc.co.uk/news/uk-53810655

• https://www.bbc.co.uk/news/explainers-53807730

• https://www.theguardian.com/education/2020/aug/16/a-level-student-launches-legal-bid-against-ofqual

• https://unherd.com/2020/08/how-ofqual-failed-the-algorithm-test/

• https://www.telegraph.co.uk/education-and-careers/2020/08/20/ugly-truth-exams-ofqual-has-used-algorithm-since-2011-resigned/

• https://www.theguardian.com/education/2020/aug/07/a-level-result-predictions-to-be-downgraded-england

• https://www.dailymail.co.uk/news/article-8623713/Could-understand-level-grades-algorithm.html

• https://schoolsweek.co.uk/revealed-ofqual-warned-of-algorithm-legal-risk/

• https://www.wired.co.uk/article/alevel-exam-algorithm

• https://www.adalovelaceinstitute.org/blog/can-algorithms-ever-make-the-grade/

• https://www.independent.co.uk/news/education/education-news/ofqual-exam-results-algorithm-b1721658.html

• https://inews.co.uk/news/education/a-level-algorithm-what-ofqual-grades-how-work-results-2020-explained-581250

• https://www.thersa.org/blog/2020/08/exam-results-algorithm