The Social Mobility Commission builds university league tables with eight year old data
David Kernohan is Deputy Editor of Wonkhe
Tags
Who remembers the HESA Widening Participation performance indicator?
The changing whims of regulators mean that the last publication of this UKPI data took place in 2020-21. It was an attempt to provide a comparable UK-wide approach to understanding institutional progress in widening access and participation, and it used a range of indicators to identify disadvantaged students.
These included:
- Whether or not an individual attended a state school or college (where a valid UKPRN is provided)
- Whether or not they came from a low participation neighbourhood, meaning the lowest quintile as measured via the most recent available POLAR data
- Their socioeconomic background (though this was discontinued from 2016 onwards because of “low quality data”)
These were then measured against a location adjusted benchmark, which took account of where students entering a particular university hailed from, along with their subject of study and entry qualifications.
Now the reason I bring this up is because the Social Mobility Commission has decided to re-run it for the cohort which (generally) entered in 2018-19. Re-run is perhaps overegging it: we still use the state school marker, plus IDACI (an area based measure of childhood deprivation), and free school meals status (which seems to have become a marker of “working class” status). The location-aware benchmarking approach also remains.
If you want to play along at home I have a plot of the data on Wonkhe from back in the day.
While in UKPI days entry requirements only featured in benchmarks, the SMC have taken a different route – we can compare the proportion of disadvantaged young people at a provider with the proportion in that cohort overall, and the proportion in the part of that cohort that met that university’s entry requirements (with this comparison rather grandly tagged as a “social mobility coefficient”).
The SMC have chosen to express its findings as a series of league tables, admirably timed to make headlines in A level results week.
If you are familiar with sector data I can imagine that you are screaming at your screen right now, which is frankly what I did when I first read the report.
Ok, deep breaths. Why 2018-19 entry? It is entirely possible to administratively match the National Pupil Data based with HESA data right up to 2024-25 – and a later year would give a clearer and fairer picture of the state of play now. Even LEO data (which is necessarily lagged to get a picture of what graduates do one year after graduation) is available for later years than that. And the cited reason of pandemic-era mark instability doesn’t cut it. The actual regulator uses that same data to do actual regulating (apparently)
Why not use UCAS data directly (or my versions)? UCAS captures state school markers, FSM markers, and postcode (which you could use to derive IDACI if you felt like the already published IMD wasn’t hitting the mark as an area-based measure of deprivation). And you could do this for 18 year old admissions for the 2025 entry cycle, which feels far more useful.
Why not use the OfS Access and Participation data dashboard? Or, if the reason is because it is a gargantuan pain to use, my derived dashboards (which already allow you to build league tables if that happens to be your kink). This data source offers you FSM, IMD, and all kinds of other fun markers (ethnicity and gender are always popular ones). And more importantly, it allows you to take a view on social mobility that extends through continuation, completion, and attainment at university through to progression to that graduate job (the latter being the current interest in social mobility research and evaluation).
(As a personal aside, why not publish UKPRN codes alongside the data, allowing people to make the kinds of comparisons I have described above? The struggle is real.)
The conclusions drawn by SMC beg another question. The kind of thing they seem to want to say is that we should return to a pre-OfS measure of access and participation by raw numbers, and ignore all the evidence and evaluation that suggests that encouraging universities to partner with schools more generally is a much more effective means of driving a much broader conception of social mobility (of a definition previously argued for by, er, the SMC). Perhaps a chat with TASO would help?
So, if you want to say something about the fact that suitably qualified young people are choosing not to go to selective universities – why ignore the large amount of work and data collection that already maps this problem spaces in a far more nuanced, and current, way?
These rankings are going to turn up in newspapers, shorn of context or analysis, and they are going to annoy almost everyone who works to actually improve social mobility.
It is indeed a shame they couldn’t use more up-to-date data – I understand it is due to the time it takes as a government ALB to get the report approved once it was written. That data will be updated. More importantly, their methodology uniquely shines a light on a “Missing Middle” which is often ignored.
University initiatives often focus on the disadvantaged groups (often the “bottom 20%”) facing the greatest barriers. The SMC analysis shows the top 10% dominate places at high tariff universities far beyond that which academic attainment alone would suggest. The result is that the 70% in the middle – neither the most-advantaged or disadvantaged by common definition – also end up being significantly under-represented.
This missing middle is ignored.
They go to typical state schools.
They don’t have the support that the most privileged get, nor do they benefit from many university initiatives to improve access.
The concept of a social mobility coefficient and curve based on the income inequality equivalents of the Gini coefficient and Lorenz curve takes into account students across all ten SES deciles, rather than assuming disadvantage or privilege is binary.
Take the University of Oxford, where the top SES decile take about 44% of places, even though they account for only 26% of those getting the required grades. The red line in the chart in the report being progressively steeper than the blue line shows that EVERY other SES decile has more students getting the required grades than they do securing places.
Under-representation at Oxford (and other high tariff universities) isn’t confined to a least-advantaged group. The 90% majority, each of the bottom 9 SES deciles, are ALL under-represented because of the elites’ dominance.
It’s not just universities and employers where this is happening. In state schools, I saw the focus on extra support for FSM/Pupil Premium. It’s not that this isn’t needed, but it’s the impact on those in the middle, particularly those just above whatever arbitrary disadvantage definition is used.
How do we address this? Let’s treat students based on their individual circumstances, not through a binary label of “disadvantaged” “FSM” or “WP student.” Every student counts and deserves support to fulfil their potential.