Conditional unconditional offers aren’t just a moral hazard – they have a demonstrable and significant impact on the continuation and attainment of students.
That’s the argument the Office for Students would clearly like to make. As we saw back in 2020 with the first phase of this project, this isn’t quite as cut and dried as some might like. I spotted at the time that the analysis offered a timely – if unplanned – critique of the dangers of admission via clearing.
As things turned out – the offer-making question is moot. Conditional unconditional offers were formally banned during the pandemic, and are now non grata via agreement with a Universities UK statement. For these reasons, this is probably the last version of such analysis. But what is to come?
Topping up the modelling and tracking with two years more data complicates matters even further. We get data from 2018-19 and 2019-20 entry cohorts, with the latter we can reasonably expect to see the impact of Covid-19 on continuation (against many expectations at the time, the pandemic increased continuation rates by just under a percentage point). You may also recall 2020 was the historic peak for offers with an unconditional component – good news, perhaps, for applicants facing a very uncertain summer.
Better administrative data linking means we can track more students, and we have more data on predicted entry qualifications. POLAR has been replaced by TUNDRA (generally seen as a more reliable guide to mainstream participation from the state school system) – and free school meals status, care experience, and socio-economic classification.
So what do we see? There is still a small, but statistically significant, difference in progression rates for students entering with A levels and an offer with unconditional components – this has decreased in the last two years to between 0.0 and 0.6 percentage points when compared to those entering with a “traditional” conditional offer.
Looking at A level entrants year by year, we see statistically significant differences between conditional and conditional unconditional offers (those ones where the offer became unconditional when you accepted it) in four out of five years – but not the most recent year.
There is no longer, by contrast, any statistically significant difference in continuation by offer type for those entering with BTECs.
If we think about entry routes, we once again see “other entry routes” roughly on a par for continuation with unconditional offers (for all qualification types) – with conditional offers one or two percentage points better than both.
Controlling for variables
We should prick up our ears a little at the findings from the predictive model. This controls for predicted entry grades, and finds a very small but statistically significant difference in continuation rates – reducing each year despite the observed growth in such offers – between conditional and unconditional offer holders for entrants with A levels.
And here we see our policy pivot:
This would be consistent with what would be found if poorer performance at A-level relative to predicted grades, for those placed through unconditional offers, were driving the lower continuation rates of these entrants.
Now. Consider the DfE interest in minimum student finance eligibility requirements. What OfS is implying here is that this research presents a great case for MERs… why should we admit applicants with poor A levels if they are more likely to drop out of their studies before they get the qualification they wanted. There’s even a handy link to the evidence on the access and participation plan data dashboard.
We also get a paragraph on a well-controlled (for difference between actual and predicted grades, level 2 attainment, and other factors) 2019 UCAS study that found a more significant impact for unconditional offers’ impact on achieved grades.
What’s bad now?
So OfS hedges bets a little with stuff like this chart:
For 2019-20 entrants, the impact of unconditional offers is more notable for those with lower predicted grades. These are small samples so we hear nothing about statistical significance – and we also see counter examples like students with really good BTEC performance (D*,D*,D*) doing better with an unconditional offer than otherwise.
We also get estimates on some of the controlling variables:
For example, in 2019-20, the estimated difference in continuation rates, between entrants reporting a mental health condition and those reporting no disability is between -1.7 and -2.7 percentage points; and the estimated difference between men and women is between -0.9 and -1.3 percentage points; and the estimated difference between FSM entrants and non-FSM entrants is between -0.9 and -1.5.
Suggesting that these other factors may be having more of an impact.
It also turns out that actual unconditional offers are a lot worse for continuation rates of A level entrants than their conditional unconditional brethren – the latter being roughly as bad as entering HE via direct clearing or application outside of UCAS.
What if the real unconditional offers were the friends we made along the way?
It’s the start of a very good recycling job – I expect future modified iterations of this work to focus on the continuations of students with less impressive entry qualifications instead. Almost as if having solved one problem at the behest of a moral panic it is time to move on to the next one.
We are iterating around a conclusion that was a commonplace 10 years ago – that students from non-traditional backgrounds need more support to flourish in higher education. Though these students enter via unconventional routes and without the constellations of starred grades that others are awarded, the ability to successfully move from getting in to getting on is there – but it is expensive to do, and background factors will also complicate outcomes data.
Table B2 shows the number of entrants placed through three entry routes (conditional offers, unconditional offers and ‘other UCAS routes’) and the continuation rates of those entrants, for each predicted entry grade profile, for five academic years of entry. This doesn’t include direct to clearing or direct to provider (RPA) routes.