- By Sean Brophy (@seanbrofee), Senior Lecturer at the Centre for Decent Work and Productivity, Manchester Metropolitan University.
A persistent challenge in UK higher education is the ethnicity degree awarding gap – the difference between White and ethnic minority students receiving top degrees (firsts or 2:1s). The Office for Students (OfS) aims to entirely eliminate this gap by 2030/31, but what if most of this gap reflects success in widening participation rather than systemic barriers?
Between 2005/6 and 2021/22, university participation grew 21% faster for Asian students and 17% faster for Black students compared to White students. This remarkable success in widening access might paradoxically explain one of the UK’s most persistent higher education challenges.
Figure 1 presents ethnicity gaps over time compared to a White baseline (the grey line constant at zero). The data for 2021/22 shows significant gaps: 21 percentage points for Black students, 9 for Asian students, and 4 for Mixed ethnicity students compared to their White peers. Traditional explanations focus on structural barriers, cultural differences, and potential discrimination, and much of the awarding gap remains unexplained after adjusting for prior attainment and background characteristics. However, a simpler explanation might be hiding in plain sight: the gap may also reflect a statistical effect created by varying participation rates across ethnic groups.
Ethnicity Degree Awarding Gap (2014/15 – 2021/22)
Here is the key insight: ethnic minority groups now participate in higher education at remarkably higher rates than White students, which likely then drives some of the observed ethnicity awarding gaps. Figure 2 presents the over-representation of ethnic groups in UK higher education relative to the White reference group (again, the constant grey line). The participation gap has grown substantially – Asian students were 22 percentage points more likely to attend university than White students in 2021/22, with Black students 18 points higher.
Over-representation of ethnic groups in HE compared to White baseline (2005/6-2021/22)
This difference in participation rates creates an important statistical effect, what economists call ‘compositional effects’. When a much larger proportion of any group enters university, that group may naturally include a broader range of academic ability. Think of it like this: if mainly the top third of White students attend university, but nearly half of ethnic minority students do, we would expect to see differences in degree outcomes – even with completely fair teaching and assessment.
This principle can be illustrated using stylized ability-participation curves for representative ethnic groups in Figure 3. These curves show the theoretical distribution of academic ability for Asian, Black, and White groups, with the red shaded area representing the proportion of students from each group accepted into higher education in 2021/22. It would be surprising if there was no degree awarding gap under these conditions!
Stylized ability-participation curves by ethnic group
This hypothesis suggests the degree awarding gap might largely reflect the success of widening participation policies. Compositional effects like these are difficult to control for in studies, and it is noteworthy that, to date, no studies on the ethnicity awarding gap have adequately controlled for these effects (including one of my recent studies).
While this theory may offer a compelling statistical explanation, future research pursuing this line of inquiry needs to go beyond simply controlling for prior achievement. We need to examine both how individual attainment evolves from early education to university, using richer measures than previous studies, and how the expansion of university participation has changed the composition of student ability over time. This analysis must also account for differences within broad ethnic categories (British Indian students, for example, show different patterns from other Asian groups) and consider how university and subject choices vary across groups.
My argument is not that compositional effects explain everything — rather, understanding their magnitude is crucial for correctly attributing how much of the gap is driven by traditional explanations, such as prior attainment, background characteristics, structural barriers, or discrimination. Only with this fuller picture can we properly target resources and interventions where they’re most needed.
If this hypothesis is proven correct, however, it underscores why the current policy focus on entirely eliminating gaps through teaching quality or support services, while well-intentioned, may be misguided. If gaps are the statistically inevitable result of differing participation patterns among ethnic groups, then institutional interventions cannot entirely eliminate them. This doesn’t mean universities shouldn’t strive to support all students effectively – but it does require us to fundamentally rethink how we measure and address educational disparities.
Rather than treating all gaps as problems to be eliminated, we should:
- Fund research which better accounts for these compositional effects.
- Develop benchmarks that account for participation rates when measuring degree outcomes.
- Contextualize the success of widening participation with acknowledging awarding gaps as an inevitable statistical consequence.
- Focus resources on early academic support for students from all backgrounds who might need additional help, particularly in early childhood.
- Explore barriers in other post-16 or post-18 pathways that may be contributing to the over-representation of some groups in higher education.
Thank you Sean for your blog. I have two questions for you: 1) If the theory that participation rate impacts the awarding gap, why don’t we see this with gender? Women are vastly over represented in HE, but achieve more higher than their male peers, not less. 2) Does controlling for prior attainment not cover the issue you outline above?
Hi Rose,
Thanks for your comment. Regarding your first point, males slightly outperform females in achieving first-class degrees, which supports the hypothesis that compositional effects matter—i.e., a higher proportion of females means a broader range of abilities, leading to lower first-class rates but higher 2:1 attainment. However, my hypothesis isn’t a panacea. Motivation also plays a key role, and research suggests that females tend to be more academically motivated than males. Regarding your second point, controlling for prior attainment will not adjust for these compositional effects, and I’ve addressed this point in the blogpost with links you can explore.
Has parental persuasion, family context of 1 or 2 parents and whether more learners prefr to get out earning and become mature students?
Widening Participation was often seen being for other races and not first in family to go to university forgotten when info handed out.
This can cause further barriers…
Perhaps some norm refrencing might be used to avoid lower level degree achievement.
For last 10 to 15 years too many were pushed into taking A Levels becsuse no alternstives eere eadiky accesdible, resulting in too many applying for Degrees instead of jobs or Degree Apprenticeships?
Interesting blog. Would not admissions criteria control for some of this compositional effect at institution level?
If e.g. all students still need AAA to be admitted, and we still see an ethnicity award gap, then it seems immaterial at the level of an institution that more students from ethnically minoritised groups are attending.
Interesting piece. The logic of the argument is that as the participation rates of black / Asian / mixed students increases relative to white students, so we should expect awarding gaps also to increase. But doesn’t the data actually indicate the opposite, i.e. that awarding gaps have (slightly) narrowed, while participation gaps have widened.
I suppose we could interpret this to mean that University efforts to reduce inequalities have been more effective than we have given them credit for. Or perhaps it means that the assumed relationship between awarding gaps and participation gaps is more assumed than real?
Whether we agree with this that the awarding gap is due to population over-sampling depends on the assumptions we make about why we’re under-sampling white and over-sampling black/Asian applicants. In an all-equal world, that ought to do nothing to awarding gaps, because we would oversample at both ends of the distribution (high as well as low academic ability students).
So other assumptions are introduced in this account. The first is that, as universities, we are selecting students from the population based on level 3 ability, but *lowering* the bar for black and Asian students. But admissions works in a very different way, because students apply for university, rather than universities selecting them from the population. Because of this, admissions are based a much wider range of factors – applicants’ abilities, motivations, beliefs, career goals etc. etc., which all likely vary according to race (e.g., the well-established disengagement of white young men from HE), and all of which likely determine performance in the degree programme in addition to academic ability.
The argument here is that we get worse academic ability in black and Asian students because we pick more from these groups on an even playing field of ability. But it could equally be (and very likely is) that we get greater motivations in these groups to attend HE (e.g., compared to white young men) which ought to lead to stronger outcomes on the degree and a smaller awarding gap. In reality it’s a mix of many factors, but the simplification to academic ability is misleading.
Another idea at play is that academic ability is unaffected by attending university. I think we have to dispute this or what are universities up to? It also seems to be assumed that level 3 performance (the basis of selection) is a context-free indication of academic ability. Agreeing with that stands in opposition to the rationale for widening participation in HE – namely that certain groups are systematically disadvantaged in obtaining level 3 quals.