2021 university admissions realities

What does mitigating pandemic disruption on the 2021 cohort mean for grades and numbers?

[This article was originally published by wonkhe.com. We have used public data from UCAS in the analysis which can be found here.]

The 2020 admission cycle was difficult. The same will be true of 2021, but in different ways. We are entering a critical period as universities begin to see the level and nature of their ‘firm’ replies as applicants approach the reply deadline in early June. A year ago, universities were facing twin uncertainties of how many students they would get and how to select fairly without exams. It is now clear that overall, the pandemic has increased, rather than reduced, demand for higher education. Fears about not having any students have receded. But the other problem, who to confirm, has grown and amplified by high demand, will be a challenge this year. The number of UK 18-year-olds seeking a university place is higher than ever before, whilst the information available for fairly choosing between these applicants is at an unprecedented low.

Three 2021 admissions realities

At the centre of the problem this year are the UK 18-year-old university applicants. This cohort has three unchangeable admissions realities caused by the pandemic.

 

First, actual learning levels are much lower than normal. Around half of their post-16 learning time has been severely disrupted, compared to around a quarter for the 2020 cohort. With these figures it is inevitable that this cohort will have learnt less of what it is post-16 study teaches. If exams were held and marked ‘as normal’, their grades would surely be substantially lower than usual. This does not mean that their potential to do well at university is reduced, but what they have learnt is. That is the reality. In the discussions of how grades are going to be awarded without exams and what the distribution is going to be, sometimes this reality gets obscured. Every A level candidate could be awarded a A*, or they could be awarded an E. But the decision makes no difference to what they have learnt. Or how well matched for university they are.

 

Secondly, the resolution in measuring learning levels is lower. Teachers will have had reduced contact time with their pupils, which introduces more uncertainty into any assessment of how people are doing. Similarly, exam-based interim assessments will generally not have been held. UCAS predicted grades last year were a stable pre-pandemic anchor, that is not the case this year. There is much greater uncertainty throughout the measurement system.

 

The third reality is that impact of these two factors will be unfairly uneven across different schools and different types of students. Which raises the question: if you could measure accurately what has been learnt over the past year, and then use this to differentiate between young people for entry to university, would you want to?

 

These three unchangeable realities that set the scene for admissions this year.

Fairness and grades

Young people who find themselves in this reality, through no fault of their own, would, ideally, be treated fairly. With ‘fair’ meaning to mitigate, as far as possible, the impact of the pandemic on their life chances. This is obviously the desire of both universities and governments. But it is not always clear what type of fairness is intended. For university admissions there are broadly three choices: to seek to admit people based on (A) their actual level of learning (which would be lower than the 2019 cohort); (B) what they probably would have got in a non-pandemic year (about the same as 2019); or (C) what they could have got in a non-pandemic year (much higher than 2019).

 

Running admissions where fairness is taken as the actual level of learning would naturally mean lower awarded grades and falling university entry rates this year. Few argue for this. Most instead say things that suggest the desire if for a ‘would have got’, or ‘could have got’, type of fairness. The difference between the two can be large for university admissions.

 

You can get a sense of this from looking at the UCAS data on the set of A level grades awarded to 18-year-old university applicants last year (Figure 1, point per grade of best three A levels, so AAA=15, AAB=14 and so on). Almost all these 2020 applicants were pre-pandemic, so if 2020 had been ‘normal’ their distribution of grades would have been quite similar to that seen in 2019 (distribution 1). The first, and short-lived, calculated grades approach in 2020 gave a distribution similar to this. It was essentially using the ‘would have got’ got idea of fairness. This view of fairness, perhaps seen as reasonable enough at aggregate level, was not accepted as fair at individual level where a ‘could have got’ perspective is more natural. Teacher-estimated grades (distribution 2) were used instead. Teacher estimates do seem to reflect the ‘could have got’ idea of fairness and so were shifted to the right, to higher grades.

Figure 1 Predicted and awarded grades for recent 18-year-old A level UCAS applicants
Figure1

In fact, they ended up somewhere between 2019 exam distribution and the 2020 UCAS predicted grade distribution (typically made December 2019/January 2020, pre-pandemic). (The UCAS data does not allow an exact like-for-like comparison between predicted and awarded grade populations, but it is close enough for this purpose). This makes sense if you view the normal UCAS predicted grades themselves as an estimate of the realistic upper limit of how well someone ‘could’ do in intrinsically random exams. See here and here for more on this reasoning. 

 

When teachers were making their 2020 exam grade assessment, they would have had a bit more information about each applicant than they did when they made the UCAS predicted grade estimates some months earlier. More information implies lower uncertainty, the range of possible outcomes narrows and so, everything else being equal, the best grades that someone ‘could’ get falls a bit.

 

In providing teacher assessed grades for the 2021 cohort teachers seem likely to seek to estimate how someone ‘could’ have done if there was no pandemic disruption. It is hard to know how they will respond to this, or how governments will seek to steer these assessments. But given the higher uncertainty in where the cohort is, it would not be unreasonable for a ‘could have’ grade distribution to shift to higher grades again, perhaps by a similar amount as it did last year. This would place the awarded grade distribution for 2021 close to the UCAS predicted grade distribution from 2020. Such an upward move in awarded grades is a necessary consequence of seeking fairness in the context of increasing uncertainty about what an individual could have got if the pandemic had not occurred. ‘Grade inflation’ perhaps is not the most helpful term in these circumstances.

 

These changes have substantial consequences for university admissions. Figure 2 shows the proportion of young A level UCAS applicants who are awarded certain grade thresholds. There were exceptionally large annual moves in the proportion awarded higher grade profiles in 2020. If the results awarded in 2021 try to fairly accommodate the increased uncertainty again, then further rises are likely. If the shift turns out to be as large as going to the 2020 distribution of predicted grades, then the proportion gaining AAB or above could rise above 50 per cent.

 

This is a particular problem for universities who exercise selection amongst higher grade students. They are not in the acute position of 2020, where pre-pandemic offer conditions were satisfied by post-pandemic grades, and many will have altered their offer making in response. But one important lesson from 2020 is that, quite naturally, the fairness expectations of applicants are set by their awarded ‘could have’ grades. If awarded grades profiles shift higher again, and placed rates (or the proportion entering higher tariff providers) does not respond then some applicants may well feel they have not been treated fairly.

Figure 2 Proportions awarded selected higher grade thresholds, 2021 distribution set to 2020 predicted grades
Figure2

Fairness and numbers

The single most important factor for mitigating the unfairness is maximising the intake.

 

Several indicators are encouraging here for young UK applicants. The headline figures show the demand increase seems more evenly spread across provider types than normal, optimising use of sector capacity. And the fall in applicants from the EU might increase the capacity for UK recruitment by 12,000 or so, particularly important in Scotland. Nevertheless, a combination of rising 18-year-olds in the population, jumps in application rates and stronger than usual demand from older UK age groups (even after allowing for the inflation effect of a later deadline this year) signal testing increases in university intakes if fairness is to be realised.

 

Figure 3 shows what different ideas of fairness for the pandemic affected UK 18 year old cohort mean for recruitment (calculated at country level, and assuming patterns of demand increases at January deadline are reflected at other stages in the cycle, and that the nature of applicants are broadly similar to last year). This is not a forecast; it takes no account of number controls, for instance, it is simply illustrating what different types of fairness mean to the numbers.

Figure 3 University entry numbers for UK 18-year-olds under different 2021 fairness models
Figure3

Given that the uncertainty of assessing potential for this cohort is greater than last year, maintaining the placed rate at the 2020 level is arguably the absolute floor of treating the 2021 cohort fairly in terms of university entry. This would give an increase on last year of some 25,000-30,000 to around 285,000. One way of fairly responding to the increased uncertainty this year is to allow the 18-year-old placed rate to increase again, say by a similar amount (3 per cent) as the UK average last year. This would boost the intake by a further 10,000 or so, perhaps reaching around 295,000.

 

That is probably getting close to the number of UK 18-year-old applicants who actually want to go to university this year. The very highest placed rates seen for UK 18-year-old A level across combinations of achieved grades and area backgrounds in 2020 were around 95 per cent. This can be seen as the level the placed rate settles at when getting the grades is not a problem (some do not take up a place because they have options outside of the UCAS system or simply want to do something else). Suppose you were to take a view that the pandemic had introduced so much uncertainty and differential outcomes to the 2021 cohort that it was not ever fair to say, ‘based on the 2021 exam results, this applicant does not have the academic potential for university’. To not turn any UK 18-year-old away on the basis of their grades would probably need a placed rate of around 95 per cent. This would result in the intake rising again, by perhaps another 10,000 or so to around 305,000.

Admissions realities

These are logistically challenging numbers. Even maintaining the placed rate from last year would make the UK 18-year-old 2021 intake some 45,000 larger, almost 20 per cent, than it was in 2019. It is far from clear that this can be accommodated after the large intakes last year, and difficult months lie ahead for applicants and universities. Governments might help here by giving universities flexibility and support, both in how many they admit and how they choose to do so. It might also be helpful to recognise that they would again be asking universities to take in students with lower learning than normal, with the intent they provide extra support to gradually repair the pandemic learning damage over the next three years. This will be costly. It is so clearly the right thing to do that universities will strive to do it in any circumstance, but governments could assist by being open to at least keeping real-terms per-capita funding levels constant.

 

The pandemic has created another exceptional admissions cycle and there are no difficulty-free routes to complete it. It will be hard to be completely fair across different types of qualifications, or those who have taken their exams in different years. And with the best of intent some universities will simply surpass even emergency levels of intake capacity before they can take everyone that fairness suggests should be given a chance. But overall, the UK HE sector continues to impress with its ability to respond to changes in demand and getting people through to graduation. If it is given the right policy environment it will likely grow intakes again, helping to minimise the long-term effects of the pandemic on young people. And in doing so when demand is higher than ever, and evidence on academic strength lower than ever, it is possible we may subsequently find that potential to benefit from university is more widely distributed than previously assumed.

[This article was originally published by wonkhe.com. We have used public data from UCAS in the analysis which can be found here.]

The 2020 admission cycle was difficult. The same will be true of 2021, but in different ways. We are entering a critical period as universities begin to see the level and nature of their ‘firm’ replies as applicants approach the reply deadline in early June. A year ago, universities were facing twin uncertainties of how many students they would get and how to select fairly without exams. It is now clear that overall, the pandemic has increased, rather than reduced, demand for higher education. Fears about not having any students have receded. But the other problem, who to confirm, has grown and amplified by high demand, will be a challenge this year. The number of UK 18-year-olds seeking a university place is higher than ever before, whilst the information available for fairly choosing between these applicants is at an unprecedented low.

Three 2021 admissions realities

At the centre of the problem this year are the UK 18-year-old university applicants. This cohort has three unchangeable admissions realities caused by the pandemic.

 

First, actual learning levels are much lower than normal. Around half of their post-16 learning time has been severely disrupted, compared to around a quarter for the 2020 cohort. With these figures it is inevitable that this cohort will have learnt less of what it is post-16 study teaches. If exams were held and marked ‘as normal’, their grades would surely be substantially lower than usual. This does not mean that their potential to do well at university is reduced, but what they have learnt is. That is the reality. In the discussions of how grades are going to be awarded without exams and what the distribution is going to be, sometimes this reality gets obscured. Every A level candidate could be awarded a A*, or they could be awarded an E. But the decision makes no difference to what they have learnt. Or how well matched for university they are.

 

Secondly, the resolution in measuring learning levels is lower. Teachers will have had reduced contact time with their pupils, which introduces more uncertainty into any assessment of how people are doing. Similarly, exam-based interim assessments will generally not have been held. UCAS predicted grades last year were a stable pre-pandemic anchor, that is not the case this year. There is much greater uncertainty throughout the measurement system.

 

The third reality is that impact of these two factors will be unfairly uneven across different schools and different types of students. Which raises the question: if you could measure accurately what has been learnt over the past year, and then use this to differentiate between young people for entry to university, would you want to?

 

These three unchangeable realities that set the scene for admissions this year.

Fairness and grades

Young people who find themselves in this reality, through no fault of their own, would, ideally, be treated fairly. With ‘fair’ meaning to mitigate, as far as possible, the impact of the pandemic on their life chances. This is obviously the desire of both universities and governments. But it is not always clear what type of fairness is intended. For university admissions there are broadly three choices: to seek to admit people based on (A) their actual level of learning (which would be lower than the 2019 cohort); (B) what they probably would have got in a non-pandemic year (about the same as 2019); or (C) what they could have got in a non-pandemic year (much higher than 2019).

 

Running admissions where fairness is taken as the actual level of learning would naturally mean lower awarded grades and falling university entry rates this year. Few argue for this. Most instead say things that suggest the desire if for a ‘would have got’, or ‘could have got’, type of fairness. The difference between the two can be large for university admissions.

 

You can get a sense of this from looking at the UCAS data on the set of A level grades awarded to 18-year-old university applicants last year (Figure 1, point per grade of best three A levels, so AAA=15, AAB=14 and so on). Almost all these 2020 applicants were pre-pandemic, so if 2020 had been ‘normal’ their distribution of grades would have been quite similar to that seen in 2019 (distribution 1). The first, and short-lived, calculated grades approach in 2020 gave a distribution similar to this. It was essentially using the ‘would have got’ got idea of fairness. This view of fairness, perhaps seen as reasonable enough at aggregate level, was not accepted as fair at individual level where a ‘could have got’ perspective is more natural. Teacher-estimated grades (distribution 2) were used instead. Teacher estimates do seem to reflect the ‘could have got’ idea of fairness and so were shifted to the right, to higher grades.

Figure 1 Predicted and awarded grades for recent 18-year-old A level UCAS applicants
Figure1

In fact, they ended up somewhere between 2019 exam distribution and the 2020 UCAS predicted grade distribution (typically made December 2019/January 2020, pre-pandemic). (The UCAS data does not allow an exact like-for-like comparison between predicted and awarded grade populations, but it is close enough for this purpose). This makes sense if you view the normal UCAS predicted grades themselves as an estimate of the realistic upper limit of how well someone ‘could’ do in intrinsically random exams. See here and here for more on this reasoning. 

 

When teachers were making their 2020 exam grade assessment, they would have had a bit more information about each applicant than they did when they made the UCAS predicted grade estimates some months earlier. More information implies lower uncertainty, the range of possible outcomes narrows and so, everything else being equal, the best grades that someone ‘could’ get falls a bit.

 

In providing teacher assessed grades for the 2021 cohort teachers seem likely to seek to estimate how someone ‘could’ have done if there was no pandemic disruption. It is hard to know how they will respond to this, or how governments will seek to steer these assessments. But given the higher uncertainty in where the cohort is, it would not be unreasonable for a ‘could have’ grade distribution to shift to higher grades again, perhaps by a similar amount as it did last year. This would place the awarded grade distribution for 2021 close to the UCAS predicted grade distribution from 2020. Such an upward move in awarded grades is a necessary consequence of seeking fairness in the context of increasing uncertainty about what an individual could have got if the pandemic had not occurred. ‘Grade inflation’ perhaps is not the most helpful term in these circumstances.

 

These changes have substantial consequences for university admissions. Figure 2 shows the proportion of young A level UCAS applicants who are awarded certain grade thresholds. There were exceptionally large annual moves in the proportion awarded higher grade profiles in 2020. If the results awarded in 2021 try to fairly accommodate the increased uncertainty again, then further rises are likely. If the shift turns out to be as large as going to the 2020 distribution of predicted grades, then the proportion gaining AAB or above could rise above 50 per cent.

 

This is a particular problem for universities who exercise selection amongst higher grade students. They are not in the acute position of 2020, where pre-pandemic offer conditions were satisfied by post-pandemic grades, and many will have altered their offer making in response. But one important lesson from 2020 is that, quite naturally, the fairness expectations of applicants are set by their awarded ‘could have’ grades. If awarded grades profiles shift higher again, and placed rates (or the proportion entering higher tariff providers) does not respond then some applicants may well feel they have not been treated fairly.

Figure 2 Proportions awarded selected higher grade thresholds, 2021 distribution set to 2020 predicted grades
Figure2

Fairness and numbers

The single most important factor for mitigating the unfairness is maximising the intake.

 

Several indicators are encouraging here for young UK applicants. The headline figures show the demand increase seems more evenly spread across provider types than normal, optimising use of sector capacity. And the fall in applicants from the EU might increase the capacity for UK recruitment by 12,000 or so, particularly important in Scotland. Nevertheless, a combination of rising 18-year-olds in the population, jumps in application rates and stronger than usual demand from older UK age groups (even after allowing for the inflation effect of a later deadline this year) signal testing increases in university intakes if fairness is to be realised.

 

Figure 3 shows what different ideas of fairness for the pandemic affected UK 18 year old cohort mean for recruitment (calculated at country level, and assuming patterns of demand increases at January deadline are reflected at other stages in the cycle, and that the nature of applicants are broadly similar to last year). This is not a forecast; it takes no account of number controls, for instance, it is simply illustrating what different types of fairness mean to the numbers.

Figure 3 University entry numbers for UK 18-year-olds under different 2021 fairness models
Figure3

Given that the uncertainty of assessing potential for this cohort is greater than last year, maintaining the placed rate at the 2020 level is arguably the absolute floor of treating the 2021 cohort fairly in terms of university entry. This would give an increase on last year of some 25,000-30,000 to around 285,000. One way of fairly responding to the increased uncertainty this year is to allow the 18-year-old placed rate to increase again, say by a similar amount (3 per cent) as the UK average last year. This would boost the intake by a further 10,000 or so, perhaps reaching around 295,000.

 

That is probably getting close to the number of UK 18-year-old applicants who actually want to go to university this year. The very highest placed rates seen for UK 18-year-old A level across combinations of achieved grades and area backgrounds in 2020 were around 95 per cent. This can be seen as the level the placed rate settles at when getting the grades is not a problem (some do not take up a place because they have options outside of the UCAS system or simply want to do something else). Suppose you were to take a view that the pandemic had introduced so much uncertainty and differential outcomes to the 2021 cohort that it was not ever fair to say, ‘based on the 2021 exam results, this applicant does not have the academic potential for university’. To not turn any UK 18-year-old away on the basis of their grades would probably need a placed rate of around 95 per cent. This would result in the intake rising again, by perhaps another 10,000 or so to around 305,000.

Admissions realities

These are logistically challenging numbers. Even maintaining the placed rate from last year would make the UK 18-year-old 2021 intake some 45,000 larger, almost 20 per cent, than it was in 2019. It is far from clear that this can be accommodated after the large intakes last year, and difficult months lie ahead for applicants and universities. Governments might help here by giving universities flexibility and support, both in how many they admit and how they choose to do so. It might also be helpful to recognise that they would again be asking universities to take in students with lower learning than normal, with the intent they provide extra support to gradually repair the pandemic learning damage over the next three years. This will be costly. It is so clearly the right thing to do that universities will strive to do it in any circumstance, but governments could assist by being open to at least keeping real-terms per-capita funding levels constant.

 

The pandemic has created another exceptional admissions cycle and there are no difficulty-free routes to complete it. It will be hard to be completely fair across different types of qualifications, or those who have taken their exams in different years. And with the best of intent some universities will simply surpass even emergency levels of intake capacity before they can take everyone that fairness suggests should be given a chance. But overall, the UK HE sector continues to impress with its ability to respond to changes in demand and getting people through to graduation. If it is given the right policy environment it will likely grow intakes again, helping to minimise the long-term effects of the pandemic on young people. And in doing so when demand is higher than ever, and evidence on academic strength lower than ever, it is possible we may subsequently find that potential to benefit from university is more widely distributed than previously assumed.