How much do we really know about HE entrants with low prior attainment?
David Kernohan is Deputy Editor of Wonkhe
Tags
The policy impetus to implement a minimum eligibility requirement for student finance was explicitly justified based on a finding from the Institute for Fiscal Studies report that those with lower prior attainment saw few benefits from higher learning.
I’ve been wondering why ministers chose not to use the DfE’s own analysis of differing returns based on prior attainment to justify this decision. And I’ve identified two reasons why they didn’t.
The first point is well understood: there’s an argument that educational attainment is a proxy for a lack of childhood disadvantage. Space and time to concentrate on academic work, a nice house in a good area with good schools, a climate of educational aspiration, and additional tuition either from well-educated family members or a paid tutor – these are the things, rather than innate ability, that point to the kinds of qualifications that put you on the higher educational ladder.
In 2011-12 (I know that’s a 2013-14 link, the 2011-12 data is available via the drop down on table 1 of the excel sheet download) some 70.4 per cent of disadvantaged pupils (here this refers to those eligible for free school meals at any point between year 6 and year 11, or with “looked after” status”) managed five GCSEs at grade A*-C: this compares to 87.3 per cent who did not.
If we exclude equivalent qualifications (non-GCSE level 2 qualifications, generally vocational in nature) and insist on a good pass in both English and Maths, the numbers are 29.8 per cent for disadvantaged pupils, and 59.4 for the rest. Twenty five per cent of the cohort who took GCSEs that year were classified as disadvantaged on these metrics.
There’s nothing particularly remarkable about the 2011-12 cohort, other than that these are the young people who form the basis of the population used in DfE’s first release of comparative annualised earnings outcomes by post-18 pathway. Every finding of that comparison needs to be read (as, indeed, does any and all salary data) through the lens of the interactions between disadvantage and attainment presented above.
The part of the recent LEO release that compares the salary outcomes of the part of the cohort that went on to level 6 qualifications doesn’t include the numbers of graduates in each quintile – it’s a puzzling omission, and leaves us with the impression that we may be comparing the outcomes for a very large group (quintile 1, nothing beyond level 3) with a very small group (quintile one, level 6 and above).
Happily – we have access to that information elsewhere. If you took GCSEs in 2011-12, you most likely took A levels in 2013-14 and graduated in 2016-17. The main LEO release offers data on those who entered higher education in that cohort one, three, and five years after graduation.
(a note: yes, it is the numbers of the cohort that entered HE, not the numbers that qualified. That’s how LEO works, for reasons I’ve never entirely understood).
This dashboard shows the numbers, median earnings (and on mouseover, the proportion with a sustained destination) for that cohort – by each of prior attainment quintile, free school meals flag (those who have ever been eligible for FSM, so a proxy for the school disadvantage flag), IDCAI quintile (an area-based measure of childhood deprivation), and POLAR4 quintile (an area-based measure of historic 18 year old participation rates).
The thing that should be ringing alarm bells is on the top left. How many young people in prior attainment quintile 1 (that’s the four Cs and and E people) graduated in 2016-17? That’s 215 – of which 120 have their earnings included in the calculation of our medians. And that includes 85 men, of which just 45 had their earnings included. That is a very small number of points of data to be comparing with anything using medians and quartiles: a very small number of external events (caring responsibilities, illness, and so on) could have a huge impact. Especially in a largely disadvantaged cohort that does not have family resources to fall back on.
We don’t have numbers for the other end of the comparison, but we could take a punt at their magnitude. There were 561,315 pupils in the 2011-12 GCSE cohort, of which about 465,891 (or 83 per cent) did not manage to get 5 GCSEs at grades A* to C. If we assume that this number minus our HE cohort of 215 is our non-HE population (of which maybe a proportion went on to L4+ apprenticeships) then we can see a problem.
The underlying LEO data is set out in such a way that it is not possible to recreate the higher education end of the earnings comparison directly – we don’t have the custom subject splits, and we don’t have any data for 2023-24 in the main release. But the sheer difference in the number of available data points – and the complicating factors around deprivation – suggests that it is barely even indicative of a difference in life outcomes.
And if DfE could only find 120 usable data points for those in the lowest prior attainment quintile that attended HE for the cohort that took GCSEs in 2011-12, you have to ask how many the Institute for Fiscal Studies were able to find in the data for from the class of 2001-02…
You write “Space and time to concentrate on academic work, a nice house in a good area with good schools, a climate of educational aspiration, and additional tuition either from well-educated family members or a paid tutor – these are the things, rather than innate ability, that point to the kinds of qualifications that put you on the higher educational ladder.”
In this you are suggesting that there is a difference in innate ability between people. This makes you a racist, fascist, white supremacist misogynist. Don’t you know that academic social science has shown that there is absolutely no difference in innate ability between people?
/s