Ask the Admissions Teams - Results Day 2020

Thank you for this message. We are not being given a choice about using CAG. This is not a decision made by the University, but by the government. We have not been presented with any scope for discretion on this matter. That is not to suggest that we would have chosen to ignore them if we had been given an option to - I do not know what we would have done. Certainly, to have gone against a sector-wide use of these grades would not have been fair.

I would also like to point out that in every year we have to reject and admit students based on the grades they present with. If student A in any other year was weaker in the Admissions process but achieved the typical offer, we would have no choice but to admit them. That said, we would still have discretion about student B, as we do now. Some students who were near misses last week have been admitted already and we may yet have discretion over those who remain near misses. If your son has received predicted grades this year, rather than ones obtained by exam, he should contact his College. He may be able to defer his offer to 2021 and sit exams next summer. I cannot guarantee this, but it is a possibility.

While some schools may have inflated their CAGs, some will have been unfairly and disproportionately hit by the algorithm, meaning that students have received lower grades than their teachers believed they would in their exams. These schools are much more likely to be in deprived areas, where attainment has been historically low and from which we have received fewer applicants previously. We will never know which CAGs are higher than the students would actually have achieved and which may turn out to be lower - these students will never sit A levels so we will never have that data. We have to trust that teachers have behaved professionally - if we accept as default that they have not, it undermines our trust in schools, their predictions and references in future. This is not a situation that we wish to find ourselves in.

Finally, please do not take this out on universities, and even less so on individuals. We are not the creators of this situation. We have been not been given a choice here and are working very hard to ensure the best situation for the everyone.

Thank you. I think you are a student yourself. Not sure if you got through to Cambridge, but our best wishes where ever you choose to go. You will thrive.

Once the government went with CAG the decision was taken out of the universities hands... but who's to say Student A's CAG is incorrect?
The teachers gave a lot of thought (and process) to their CAGs, much more so than predicted grades for UCAS applications.

There are good statistical reasons why assessing ability will sometimes lead to slightly higher grades than are actually achieved.

Also, as in a normal year, by this stage, all the other conditions have been met - did well enough in the interview(s), university assessment tests, STEP etc. The exam result (this year, eventually, the CAGs) is the last piece in the puzzle. It isn't When we receive your results from UCAS, we hold a meeting to rank you all again in terms of interview score and try to work out which examiners in which subjects for each exam board mark a bit more leniently than others, so those A*s and As will possibly be rejected because we don't think that applicant really submitted an exam paper that actually meets our requirements...

Thank you. It is clear to everyone that both CAG and Ofqual moderated grades have their problems, and none of the problems was caused by the university. But just as you pointed out in the third paragraph, for the moderate grades, the university might be able to identify very roughly (or just assume) students who were likely disadvantaged, i.e., those from schools in the deprived areas. This factor had been taken into account, from what was said in your previous posts. As for CAG, the university has no way to tell, and so the approach leads to greater randomness.
We all trust professionalism of teachers and schools. But is it fair to deny the effort of Ofqual to regulate the obviously inflated grades ?
I understand that the university has no choice. Could the many smart minds in Cambridge come out to tell people whether they agree or not that the algorithm used by Ofqual is truly scientifically ungrounded, or just that it is disliked for making many students/parents unhappy.

Thank you. It is clear to everyone that both CAG and Ofqual moderated grades have their problems, and none of the problems was caused by the university. But just as you pointed out in the third paragraph, for the moderate grades, the university might be able to identify very roughly (or just assume) students who were likely disadvantaged, i.e., those from schools in the deprived areas. This factor had been taken into account, from what was said in your previous posts. As for CAG, the university has no way to tell, and so the approach leads to greater randomness.
We all trust professionalism of teachers and schools. But is it fair to deny the effort of Ofqual to regulate the obviously inflated grades ?
I understand that the university has no choice. Could the many smart minds in Cambridge come out to tell people whether they agree or not that the algorithm used by Ofqual is truly scientifically ungrounded, or just that it is disliked for making many students/parents unhappy.

Where is that quote from? And was it before the CAG U-turn?

Whether or not the algorithm is statistically sensible doesn't require a genius. It just requires access to data that nobody but the government has. Ofqual declined to allow the Royal Statistics Society (or something like that) to evaluate their model, so I'm afraid it's likely we'll never know. The algorithm is clearly not terrible though, as school performance is (vaguely) consistent year-on-year.

Not a quote, just my ramblings along the lines of how the 'normal exams year's discussion of applicants' exam grades and other 'holistic' aspects might (not) go🙂. The equivalent of looking at the CAGs and deciding which applicants' teachers will have inflated their grades and therefore they should be the ones to be rejected, despite making their offer, grades-wise.

The algorithm absolutely is terrible and it's not at all clear that it's more accurate at predicting grades that would have been achieved if students had sat exams than just using CAGs. Certainly it seems very unlikely that it is better at predicting whether or not students can cope with the Cambridge course than Cambridge admissions tutors would be, given that they have access to information such as GCSEs, admissions tests results and interviews that the algorithm doesn't have (and that's without getting into contextual data).

The algorithm is not, on the basis of the actually substantiated claims of that blog post, "absolutely terrible". It's unrefined and makes some questionable statistical assumptions, but that doesn't make it terrible. It just means they could have done better for a minority of students.

Yes it is. Teachers predict grades with 50% accuracy, the algorithm is reported to have 40-75% accuracy. But the latter figure is based on ridiculous assumptions, so the true figure is certainly much lower. What evidence do you have that it would be more accurate than using CAGs?

Perhaps they could. But we do not have access to the full data and from what I've seen, students from schools you might not expect to be downgraded have been, where others you might expect to have not. And even if we did, to say that the algorithm was good or bad risks pitting us against students and making it sound like we didn't want to take some of them, which is certainly not a position we want to be in.

Thank you again. My criticism and disappointment are largely related to that I have had very high respect to Cambridge's holistic approach and detailed process on individual applications. I have heard this from my friends and neighbours and from the detailed description of the admission team in this thread (and the previous years' versions). My child's experience in the application journey also showed the best part of the whole process, as he had walked through all the way to summer pool rejection heard in the evening of Aug 14. We really appreciated and enjoyed. Just that blanket adoption of CAG seems to have greatly contaminated the nice picture that we have had in mind.
I hope that Cambridge will stay the same Cambridge in the future. And Cambridge should take courage to express her views on "right" and "wrong", even when facing "risks pitting against students". Cambridge is a highly respected educational institution. Caring students does not mean being afraid of conflicts with students.

The latter figure is based on less-than-precise assumptions, sure. Why that suggests it's "certainly much lower" is quite beyond me, and completely unsubstantiated.

Unsurprisingly there is no good/ easy/ cheap substitute for nation-wide exams*. If there was, nation-wide exams wouldn't exist, because we could save the government, schools, teachers and students a load of money, stress and administration costs.

* - Although, a part of me does wonder what 100% candid teacher predictions would look like against actual grades. Of course any kind of "official" prediction (on UCAS, or on a student's report card) or CAG submission is going to be at least slightly inflated, but I am curious what teachers would be able to predict in terms of their students success (particularly after knowing them for 2 years). I did tutoring on/off for the last few years (mainly A-Level though I also did some first year uni Chem/Physics) and I was bad at predicting their grade. I almost always underpredicted, but I think I was biased because my lessons almost always hyper-focused on what they were worst at and we would only discuss wrong answers of past papers, etc. I wonder if I was a teacher who saw a more holistic view and had a few years experience if I would be able to achieve a 80%+ pearson score against exam results.


Ultimately I think I support the decision to err on the side of caution and grade inflation. Unlike CAG-adjusted algorithm grades (which let down some students by forcing them to underperform due to historical averages), this instead is going to potentially let down students by guiding them into courses which they are not academically prepared for. Although the concern doesn't really spread to Oxbridge because most students there are so sensible and conscientious when it comes to doing work, I am concerned about students getting in for ABB+ courses who perhaps haven't even finished most of their A-Levels, let alone done the specific work and preparation that would have been required to get those grades. This goes before the other concern that perhaps 10%+ of these grades were inflated against what would have actually been achieved anyway. However, the difference is that at least these students get a chance to prove themselves on the course, which does feel better.

And ultimately these students are 18, turning 19, and entering into the part of their life where they need to be able to take responsibility for themselves and it's probably insulting to suggest that they need to be told what to do. Going to university and failing the first year is definitely not the disaster it might sound like it would be, so long as there is an appropriate reaction and follow up decisions. But these students aren't kids and I think deserve a certain amount of trust

Not to be rude, but how much experience do you have with statistics? The testing method is obviously completely invalid to anyone with relevant knowledge. It's not that some of the assumptions are slightly generous, it's that they're completely wrong. If you think that the given figures are meaningful, the onus is on you to demonstrate how that can be the case.

Can I please ask if there is any clarity with what happens with regard to CAGs and IB students? IB schools were asked to submit their assessments in a similar way to A level but, given only 4,500 out of the 175000 IB students are in the UK, the IB is unlikely to change its methodology to agree to centre assessed grades. Will students be able to demonstrate their predictions and thus meet their offer conditions in a similar way to A level students? It would seem terribly unfair fir not. Thank you.

@Peterhouse Admissions could I possibly just nudge this question, please? It seems to have got lost in the (understandable) discussion around CAG.