disparate impact of Covid-19: race and predicted grades

On 29 April 2020 the EHRC issued a warning to Ofqual that its plan to use predicted grades in place of GCSE and A level examinations in 2020 ‘could deepen the existing inequality in education and put the future of disadvantaged young people at risk if not correctly implemented’  https://www.equalityhumanrights.com/en/our-work/news/predicted-grading-during-covid-19-could-limit-young-peoples-futures). The EHRC advised Ofqual in its response to the body’s consultation on its decision to cancel a range of exams and assessments across England that there was research evidence ‘of conscious or unconscious race bias’ when predicting grades.The EHRC called for the Department for Education to issue guidance to schools on the approach to grade predictions and student ranking in order to minimise the risk of bias and for Ofqual to ‘publish a report evaluating the predicted grades process and outcomes for pupils’ and to take action if disparities related to ‘protected characteristics’ (these include race and sex).

The EHRC consultation response followed a letter sent on 2 April 2020 from 21 academics, coordinated by the Runnymede Trust, which advised that ‘students from lower socio-economic backgrounds (particularly higher attaining students from lower SES backgrounds) are more likely to have their final grades under-predicted compared to higher attaining students from more advantaged backgrounds’. The letter referred to research by Dr Gill Wyness, warning that: ‘well over a quarter of black and minority ethnic (BAME) GCSE students (including much higher proportions of Gypsy, Roma and Irish Traveller students) are on free school meals, we are concerned that this issue of under-predictions in grades for students from lower socio-economic backgrounds will impact on BAME students disproportionately. We are also aware that “teachers’ expectations of black students and their working-class peers tend to be systematically lower than warranted by their performance in class”’ (citing David Gillborn, “Racism as Policy: A Critical Race Analysis of Education Reforms in the United States and England” (2014) 78 The Educational Forum, 78:1, 26-41).

The Wyness research (Rules of the Game: Disadvantaged students and the University admissions process) showed that there was a strikingly consistent inverse relationship between socio-economic disadvantage and predicted grades:

The poorer the student, the less likely they are to benefit from over predictions and the more likely they are to have their grades under-predicted. The opposite is true for the wealthiest students. There is a close relationship between ethnicity and socio-economic disadvantage in the UK; in 2018, for example, 31.2% of White Gypsy/Roma or Traveller people were ‘long term unemployed or never worked’ and 25.3%, 24.4% and 19.7% of those of Bangladeshi, Pakistani and Arab origin respectively, by comparison with 4.8% of white people. 15.4% and 12% of people of Indian and white Irish origin respectively were in higher managerial, administrative or professional jobs, compared with 2.5% of White Gypsy/Roma or Traveller people and 4.2% and 4.9% respectively of those of Bangledeshi and mixed white/ Black Caribbean origin. The implications of discriminatory predictions are particularly stark when, by contrast with the position when examinations are actually undertaken, pupils will likely have very limited opportunity to prove the predictions wrong.

The signatories to the letter of 2 April urged the DfE to ‘[p]rovide teachers with more guidance and support on how to ensure more accurate predictions in order to reduce inconsistencies across schools and pupils’ and ‘with guidance on how to undertake Equality Impact Assessments of final grade predictions. This could involve schools disaggregating final predicted grades by protected characteristics, as well as SEND in order to monitor and reduce inconsistencies across different groups of pupils. They also called on universities ‘to consider carefully “contextualised admissions” criteria in order to ensure that students from lower socio-economic backgrounds do not lose out from the opportunity to attend more selective universities’ and on Ofsted to require that higher education institutions ‘monitor and report on offers made on the back of school predictions by ethnic and gender group to check for any bias at the point of admissions’.