Ucas introduced the concept of the multiple equality measure (MEM) in 2015. Designed to take a holistic view of equality in higher education admissions, using several dimensions rather than the more conventional single measure, we sought to bring to the sector a novel means of supporting widening access and participation. Our new report goes into the finer details of how we’ve constructed and updated the MEM, and its usefulness for the sector.
For too long, one-dimensional metrics have been used to judge progress. Should we look at where someone lives (typically using higher education participation rates by local areas, or POLAR3, a proxy), their ethnic group, their gender, their parents’ income (as indicated by free school meal entitlement), or the type of school they went to?
By looking at these characteristics in isolation, it was inevitable that anomalies in reporting would occur. For example, how could you consider the student who lived on the outskirts of a POLAR3 quintile 1 address (an area with the lowest rate of higher education participation), whose parents had a six-figure income, and were able to send them to a private school?
Meanwhile, POLAR3 quintile 3, which sits in the middle of participation rates, is particularly tricky to manage. Students in this group are often overlooked, deemed neither advantaged nor disadvantaged, yet almost 10 per cent of these students belong to subgroups such as white males who receive free school meals. These students in particular have very low likelihoods of going to university.
Our analysts have developed the multiple equality measure to provide a multidimensional assessment of equality in admissions, thereby reducing the risk of disadvantaged applicants not being identified.
The MEM uses a data-driven model on a dataset of pupils in English schools, then links these to Ucas’ data on how these pupils progressed to higher education. The model can then predict how likely each person is to go to university based on a set of variables, none of which would directly affect an individual’s suitability to study, but which in fact do.
The variables currently used are sex, ethnic group, POLAR3 quintile, school type, the government's Index of Multiple Deprivation and free school meal eligibility. From these calculations, all students can then be aggregated into groups based on their modelled chances of starting a higher education course.
MEM group one represents those least likely to enter HE (most disadvantaged), with group five marking the students most likely to enter (most advantaged). Within a single MEM group, there can be students from very different backgrounds and could include students from almost all variable groupings.
The MEM has four key benefits: it’s accurate, data-driven, applicable at an individual level and flexible. Accuracy comes from the MEM being specific to higher education because it only looks at characteristics relevant to equality in HE, and the use of statistical modelling means that several characteristics can be analysed at the same time.
The MEM is made from an individual-level dataset, so it only identifies effects that are actually present in the data. Using detailed levels of data also means that the MEM can interpret likelihoods for everyone within it. The model approach gives flexibility both in the characteristics included, and the level at which equality can be measured.
Overall results show that, by using the MEM, the gap in entry rates between the most and least advantaged students is wider than if POLAR3 were used on its own. Using MEM, people from group five are 4.7 times more likely to go to university than those in group one, compared with a ratio of 2.3 times for POLAR3 exclusively. Since 2014, the POLAR3 gap has continued to narrow, whereas it has remained static in MEM over the same period.
The HE sector is increasingly aware of the benefits of MEM. It was mentioned in the Office for Students’ guidance on access and participation plans for 2019/20, and is also referenced in its live consultation on its future approach to regulating access and participation. We are in active discussions with the OfS about how we might collaborate to determine next steps for the measure.
The MEM has been three years in development, but its flexibility means that it is constantly evolving. Although the current version only covers England, we are speaking with the other regulatory bodies across the UK to discuss expanding the measure. Notwithstanding the ongoing need for refinement, we would encourage policymakers and the HE community to adopt MEM as the primary measure of equality.
David Best is director of analysis and insights at Ucas.