This is a brief supplementary post about progression by FSM students to selective universities.
In preparing my last post, I had occasion to look again at DfE statistics on KS5 student destinations.
These experimental statistics were first published in 2012 and most recently in January 2015. To date they cover four academic years, starting with AY2009/10 and ending with AY2012/13.
Underlying data is published each year and since AY2010/11 this has included the number of FSM students admitted to different categories of selective university: the ‘top third’, Russell Group and Oxbridge.
Allowing for a health warning about potential comparability issues (see Technical Notes below) I wanted to investigate how FSM admissions to these categories had changed over the three years in question.
The numbers are set out in this embedded spreadsheet.
On closer inspection they reveal some interesting information.
Graph 1, below, shows the percentage increase between AY2010/11 and AY2012/13 for FSM and non-FSM students in each category of selective higher education.
On the face of it, this is extremely good news for fair access, since the increase in FSM admissions significantly exceeds the increase in non-FSM admissions for all three categories of selective higher education.
The increase in FSM progression to Oxbridge is exactly in line with the increase at Russell Group universities.
The improvement at ‘top third’ HEIs is some 40 percentage points lower, but these institutions are almost 10 percentage points ahead of the rate of improvement for all HE.
Over the same period non-FSM progression to Russell Group universities has increased at almost twice the rate of non-FSM progression at Oxbridge, which is only slightly ahead of the 10% or so improvement at ‘top third’ institutions.
But non-FSM progression to all higher education has actually fallen slightly over the period.
Graph 1: Percentage increase in FSM and non-FSM students attending selective HE destinations between AY2010/11 and AY2012/13 (From DfE destination statistics, underlying data)
The similarity between the FSM increases for Oxbridge and Russell Group universities may help to substantiate the improvement for the former, despite the potentially drastic impact that rounding can have on such small totals (see Technical Notes below).
On the other hand, it should not be forgotten that this radical improvement was achieved in a single year, between AY2010/11 and AY2011/12.
In the following year there was no change at all for Oxbridge, with FSM admissions stalled on 50, whereas the improvement at Russell Group universities was much more consistent, increasing by some 22% compared with AY2011/12.
Further insights can be gleaned by looking at the figures in a different way.
Graph 2 shows the percentage of total admissions to the different categories of selective higher education accounted for by FSM students – and how these have changed by academic year.
This reveals a somewhat different picture. The FSM progression rate to Oxbridge remains some two percentage points behind the rate for progression to the Russell Group as a whole (although the gap closed temporarily in AY2011/12). Whereas there has been steady improvement across the Russell Group, the FSM share fell back back at Oxbridge between AYs 2011/12 and 2012/13.
The overall improvement for all higher education has also been strong, particularly so between AYs 2011/12 and 2012/13. At ‘top third’ universities the FSM share fell back a little in 2011/12 but recovered strongly in 2012/13.
Graph 2: Percentage of admissions to Oxbridge, RG, Top third and all HEIs accounted for by FSM students, 2010/11 to 2012/13 (From DfE destination statistics, underlying data)
One might normally be wary of expressing changes in comparatively small percentages as percentages themselves, but since the UCAS End of Cycle Report (see below) includes such calculations, it seems equally justifiable in this context.
They reveal a substantial 24-point difference in the change in the FSM share of total admissions between 2012 and 2013, with Oxbridge recording -10% and the remainder of the Russell Group +14%
This coincides with a change in the constitution of the Russell Group, as Durham, Exeter, Queen Mary’s and York Universities joined in 2012. This might have had some small impact on share, but does not explain the 24-point gap.
A more tantalising question is the impact of the relaxation of student number controls for students with A level grades of AAB+ or equivalent, combined with a fall in the total number of applicants. Did these factors contribute to the improvement at Russell Group universities, or was the improvement achieved in spite of them?
UCAS End of Cycle Data
This data provides a more differentiated view of FSM progression to selective universities than the oft-quoted UCAS End of Cycle Report 2014, which has a small section on this topic, based on matched NPD and UCAS admissions data.
FSM eligibility is determined when the student is aged 15 and selective ‘high-tariff’ institutions appear to be calculated on the same basis as the ‘top third’. This ensures a degree of comparability with the Destinations statistics, although the UCAS data relates to the progression of 18 year-olds from state-funded schools only (so excludes colleges).
Furthermore there is no expectation of sustained participation (see technical notes below) and the ‘top third’ of universities has probably been calculated in a different year.
The UCAS analysis is confined exclusively to entry rates – the proportions of the total FSM and non-FSM 18 year-old populations progressing to high-, medium- and low-tariff universities respectively.
Graph 3, below, is derived from the data underpinning the Report. It shows progression to high-tariff universities for FSM and non-FSM students.
Graph 3: FSM and non-FSM entry rates to UCAS high-tariff universities, 2011-2014
This reveals that:
- There were very small increases in entry rates between 2013 and 2014, for both FSM and non-FSM populations. (The Report notes that this is a 3.7% improvement for FSM and a 2.9% improvement for non-FSM.)
- The ratio between non-FSM and FSM has also narrowed minimally, but the gap between them has widened minimally too (from 6.4 points to 6.5 points).
- Since 2011, the FSM entry rate has increased by some 50% while the improvement in the non-FSM entry rate is nearer 25%. The ratio between the two rates has improved, but the gap between them has widened from 5.6 points to 6.5 points.
This is not the universally positive story for fair access suggested in media coverage and subsequent political commentary.
Data published by Oxford and Cambridge, either in their access agreements or admissions statistics, show that progress over the three years in question has been inconsistent.
- At Oxford the total number of applicants from Acorn 4 and 5 postcodes reached a peak of 1,246 in 2010/11, only to fall to 1,079 in 2011/12 and 1,070 in 2012/13. The percentage of all students admitted with Acorn 4 and 5 postcodes was 7.6% in 2010/11, but fell to 6.7% in 2011/12, increasing only slightly to 6.8% in 2012/13.
- At Cambridge 4.1% of applicants in 2010/11 were home applicants from Polar 2 quintile 1 postcodes and 17.6% were successful applicants. There was an improvement in 2011/12, to 4.6% of applicants and a 22.6% success rate but, in 2012/13, applications remained at 4.6% and the success rate fell back to 20.2%.
Unfortunately neither chooses to make public any data they might hold on annual admissions from FSM and non-FSM students.
Reasons cited in access agreements include the effects of the new student funding regime, a fall in the number of school leavers and the argument that an impact will only become apparent after sustained activity over a five year period. Oxford is however predicting significant improvement in AY2013/14 on the basis of its provisional data.
But one might reasonably expect these factors to have had a similar effect on other Russell Group universities. So how does one justify the disparity revealed by graph 2 above – between Oxbridge and the remainder of the Russell Group?
Possible reasons for the disparity between Oxbridge and other Russell Group universities
The explanation most often supplied by Oxbridge is that very few FSM-eligible students manage the exceptionally high attainment required for admission.
Admissions statistics from the two universities shows that, in 2012/13:
- At Oxford 37.1% of students accepted had A*A*A*, 27.2% had A*A*A, 24% had A*AA and 9.4% had AAA (best three A levels).
- At Cambridge, 59.5% of applicants achieving a UCAS tariff equivalent to A*A*A* were accepted, as were 23.6% of those with A*A*A and 13.9% of those with A*AA.
Data on FSM achievement at the highest A level grades (or equivalent) is particularly hard to come by. I have previously drawn on answers to various Parliamentary Questions that show an increase of some 45% in FSM students achieving AAA or better at A level between 2006 and 2011.
The most recent of these (Col 35W) was answered in July 2012. It says that, of pupils entering at least one A level in 2010/11 and eligible for FSM at the end of Year 11, there were 546 who achieved 3 or more GCE A levels at A*-A. This includes students in both the school and FE sectors. By comparison, there were 22,353 non-FSM students achieveing the same feat.
If we look at the ratio between achievement at this level and admission to Oxbridge in the same year:
- 546 FSM students corresponded with 30 places secured (ratio 18:1)
- 22,353 non-FSM students corresponded with 2,260 places secured (ratio 10:1)
So what exactly is happening? There are several possible further reasons for FSM under-representation:
- Too few FSM students are gaining A* grades (or equivalent), as opposed to A grades, at A level.
- Too few FSM students are gaining the necessary grades in suitable subject combinations and/or in facilitating subjects. (There has been some suggestion recently that subject choice is an issue, though this study adopts a broader definition of disadvantage and does not apply specifically to Oxbridge admission.)
- When Oxbridge chooses FSM students pre A-level, their GCSE/AS level performance does not reflect their eventual A level performance.
- Too few of the highest attaining FSM students are applying to Oxbridge, quite possibly for a variety of different reasons.
- Too many FSM applicants to Oxbridge are seeking entry to the most competitive courses; too few to those where there are fewer applicants per place. (At Oxford in 2012/13, for example, the success rate for medicine was 10% while for classics it was 42%)
- FSM students do apply in proportion, but are relatively less successful at gaining admission for reasons other than (predicted) attainment. One reason might be that neither University specifically targets FSM students through its access strategy, preferring alternative indicators of disadvantage.
Unfortunately, there is very little data available publicly to test which of these hypotheses are correct, their relative impact and how they operate in combination.
As attention switches to the pupil premium measure, one wonders whether the next government will ensure that reliable data can be made available to selective universities and, through Offa, expect them to feature this in their access targets, as well as their policies for contexualised admissions.
There is a timelag associated with the HESA dataset, which has to be matched with the National Pupil Database. For example, the January 2015 publication matches data on students in KS5 taking A level and equivalent qualifications in AY2011/12 and on those in HE in AY2012/13.
The most recent publication appeared in January 2015. Since HESA collects data at the end of each academic year the lag was approximately 18 months.
The next publication, relating to academic year 2013/14, is not scheduled for release until October/November 2015, indicating a lag of 15/16 months.
According to the Technical Note linked to the most recent SFR KS5 students are included if they:
- Entered for at least one A level or equivalent level 3 qualification similar in size to an A level.
- Attend state-funded mainstream schools, independent schools, FE and sixth form colleges and maintained, non-maintained and independent special schools. (However, it seems that only a few independent schools – those that provide tracking information to local authorities – are included.)
Students must record sustained participation – in all of the first two terms of the year – at one or more HE destinations. In 2012/13 this was defined as between October 2012 and March 2013.
Higher education is defined as any UK HE institution, so those admitted to institutions abroad are excluded. Students undertaking HE courses at FE institutions are included. The note is not quite clear about the treatment of students accepted for deferred entry.
The categories of selective HE are nested within each other:
- The top third of HEIs when grouped by mean UCAS tariff score from entrants’ best three A level grades. KS5 students with other qualifications are excluded from the calculation. For the purposes of this publication, students with no A level points were excluded from the calculation. The ‘top third’ methodology is preferred by BIS. The constitution of the group changes annually, though 88% of institutions were within scope for six consecutive years up to 2011/12. (The 2011/12 list is used on this occasion.)
- The Russell Group (Birmingham, Bristol, Cambridge, Cardiff, Durham, Edinburgh, Exeter, Glasgow, Imperial, KCL, Leeds, Liverpool, LSE, Manchester, Newcastle, Nottingham, Oxford, Queen Mary’s, Queens Belfast, Sheffield, Southampton, UCL, Warwick and York).
- Oxbridge (Oxford and Cambridge)
Eligibility for free school meals (FSM) means students eligible for and claiming FSM in Year 11. Pupil premium was not introduced until September 2011, when these students were already beyond Year 11.
All national figures are rounded to the nearest 10, which makes small totals particularly unreliable. (For example, 40 + 10 could represent 35 + 5 or 44 +14, so anywhere between 40 and 58.)
The technical note advises that:
‘Some of the differences across years may be attributable to the tightening of methodology or the improvements in data matching, so comparisons across years must be treated with caution.’