Exam Grades
Comments
-
I’ve always thought that maths is the best 😉0
-
Absolutely! I’ve always said to my (English graduate) daughter that I can still pick up and read a book but she’s got no hope solving a system of partial differential equationsmorstar said:
Well yes. All the others are simply applied maths.johngti said:I’ve always thought that maths is the best 😉
0 -
The other staggering fact is how many students still take University courses which are not very helpful in a tough job market... the number of enquires for things like "Liberal Arts" or "Italian with Film Studies" is unbelievable...left the forum March 20230
-
Having just been through choosing a secondary school, numbers of A-C (can't remember the new GCSE number equivalent) grades is very much used as a marketing tool, despite the attempt to try to move to the more useful 'value added' score. There's also the issue of different exam boards. Way back when I did my GCSEs I know that my school specifically chose an easier board to improve their pass rate. Not unreasonable as that was how schools were assessed then.john80 said:
What i have proposed does not stop you comparing schools as the underlying grade still stands. It is just the banding that moves yo keep the standard consistent across the nations. You can still assess poorly performing schools. It just means you dont have to create a new A++++++ grade as every year more kids are deemed worthy of an A. Why should say anymore than 20% of kids get the top grade.rjsterry said:
If all you were aiming to do with grades was rank the 660,000 students in order of ability to sit exams then that would work. However, the grades are also used to assess the performance of schools and individual teachers. If we accept that some schools need to improve then under the zero inflation model, others will appear to get worse as their students are bumped down the ranking. On top of this, if grades are primarily based on final exams, then the teaching will inevitably tend towards the skill of passing a 3 hour written exam at the expense of other knowledge and skills. As an employer, I'm not sure great exam technique is a useful transferable skill.john80 said:I have never understood why we allowed grade inflation as it devalues the exams. If you took the stats from 20 years ago and essentially mirrore the percentage for each grade by manipulating the yearly grade percentage this woukd be a better system. If 60% makes the right number get the top mark then the exam is too easy. If 98% gets you the top grade tbe exam is too hard. This obviously assumes that you want to differentiate the bright worthy of further education to the dim that should go into the world of work and not give everyone a feel good factor of higher grades.
Ofqual suspended their appeal criteria overnight, pending a review, so now nobody knows where they stand and a repeat is incoming with the GCSE grades.
If grades are to be based on a national ranking system, whenever a student moves up a place in the ranking, some others will move down. Nothing wrong with that on the face of it, but you can't then determine the thresholds for different grades before the results are in, only give an approximate guide based on previous years. That then makes predicted grades more unreliable.1985 Mercian King of Mercia - work in progress (Hah! Who am I kidding?)
Pinnacle Monzonite
Part of the anti-growth coalition0 -
There are now 9 grades at GCSE. Why do 11% of pupils not fall into each grade?
Likewise, if there are A-E + F grades at A-level, why aren't 1/6th of student in each grade?
Surely the point is to see where the individual lies in relation to their peer group and this means no grade inflation, just an easier assessment for those looking in from the outside as to where a person with say grades from 2015 compares with someone with grades from 2019.1 -
This is entirely moot as *the students haven’t sat the exams*Dorset_Boy said:There are now 9 grades at GCSE. Why do 11% of pupils not fall into each grade?
Likewise, if there are A-E + F grades at A-level, why aren't 1/6th of student in each grade?
Surely the point is to see where the individual lies in relation to their peer group and this means no grade inflation, just an easier assessment for those looking in from the outside as to where a person with say grades from 2015 compares with someone with grades from 2019.
The issue here is the algorithm produces grades that do not reflect the individual student’s ability. That is obviously stupid.
And this is broadly done anyway.
Once the marks are in each exam board for each paper sits down in a big meeting and decides the grade “boundaries”
You also need to recognise that grades are also used to measure school performance and improvement, so if schools are on the whole improving you should expect to see some grade inflation.
If it’s purely based on distribution how would you look to measure improvement?0 -
Is that not just potentially hiding any changes in the standards of exams? The results of each cohort would look exactly the same regardless of whether the standard of teaching or the difficulty of the exam went up or down.Dorset_Boy said:There are now 9 grades at GCSE. Why do 11% of pupils not fall into each grade?
Likewise, if there are A-E + F grades at A-level, why aren't 1/6th of student in each grade?
Surely the point is to see where the individual lies in relation to their peer group and this means no grade inflation, just an easier assessment for those looking in from the outside as to where a person with say grades from 2015 compares with someone with grades from 2019.1985 Mercian King of Mercia - work in progress (Hah! Who am I kidding?)
Pinnacle Monzonite
Part of the anti-growth coalition0 -
Rick, you've missed my point. I'm not referring to this year, hence I used 2015 and 2019. This year is a one off clearly and there are issues around the algorithim. But looking at the wider picture, as an employer, or a university, surely you want to know that someone with say a 9 at GCSE falls within the top 11% in the country in their year and that subject. How hard or easy the exam is in comparison to other years doesn't matter if you assume the most able will (almost) always do best anyway.
The current system means you have no idea whatsoever what a 9 actually means. The grade inflation of the last 25+ years just devalues the higher grades.0 -
Well the exams are supposed to be the same level of difficulty between the years so an A in 2015 is as difficult as an A in 20190
-
That might be the objective but in reality it's not the case, hence adjustments at the grade boundaries each year under the current system. Additionally someone earlier suggested that exams from a number of years ago were of a different standard.rick_chasey said:Well the exams are supposed to be the same level of difficulty between the years so an A in 2015 is as difficult as an A in 2019
0 -
It is impossible to get exactly the same difficulty exam hence the boundary tweaks.Dorset_Boy said:
That might be the objective but in reality it's not the case, hence adjustments at the grade boundaries each year under the current system. Additionally someone earlier suggested that exams from a number of years ago were of a different standard.rick_chasey said:Well the exams are supposed to be the same level of difficulty between the years so an A in 2015 is as difficult as an A in 2019
They are specifically designed to be comparable between years.
Kids are just better educated than they used to be.0 -
The 9-1 GCSEs were designed to be more academically challenging compared to the old A*-G system. Consequently the way that grades are allocated has changed significantly. The new grade 9 is there to identify exceptional performance , an A** if you like. That grade is awarded to an average of 5% of students, depending on the size of the cohort. So maths awards approximately 3.5% grade 9 over the whole cohort whereas a subject like Latin should get significantly more because it’s a minor subject and the students who do it are generally more able academically.Dorset_Boy said:Rick, you've missed my point. I'm not referring to this year, hence I used 2015 and 2019. This year is a one off clearly and there are issues around the algorithim. But looking at the wider picture, as an employer, or a university, surely you want to know that someone with say a 9 at GCSE falls within the top 11% in the country in their year and that subject. How hard or easy the exam is in comparison to other years doesn't matter if you assume the most able will (almost) always do best anyway.
The current system means you have no idea whatsoever what a 9 actually means. The grade inflation of the last 25+ years just devalues the higher grades.
The majority of grades will be between 4 and 8 but they follow a normal distribution so you’d see more 4-7 than 8/9 and 1-3 grades awarded.
Standards were set for the first year on equivalent outcomes so the boards assumed that ability of the cohort didn’t change between the last year of A*-G and the first year of 9-1 and grade boundaries were set accordingly to ensure consistency. Since then, a sample of students take a reference test each year to compare the ability of each cohort to determine where grade boundaries are put for the summer exams. This happens around February/March time and schools are chosen at random to take part (my school was involved this year). Schools don’t get to find out how students did in this.
Ultimately the harder exams (and they are certainly harder) have grade boundaries that seem a lot lower. It’s not unknown to have to get 55% overall in maths to secure grade 7, the equivalent of the old A grade, but this is set using the assumption that ability doesn’t change much, and is measured anyway via the reference test, and proportions achieving a particular grade don’t vary by much. There’s variation each year anyway because borderline grades are looked at in examiner meetings but that’s just statistical noise.
Grade “inflation” happened when teachers became more used to the requirements of the exams. If you know what’s coming then you can prepare students better for it. The old GCSE was a criterion referenced system so, crudely speaking, if you jumped through the right hoops you’d get the corresponding grade. Some argue that this is a better system because you’re rewarded for what you can do but it’s imperfect because there still has to be a degree of ranking within the cohort. The 9-1 system should avoid this because it’s more of a norm referenced system so very much based on a normal distribution.
So, basically, you do know what a 9 means. It’s that the student achieving it is in the top few percentage points of the whole cohort and is therefore very good at taking exams. Which is all this measures, really.0 -
A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%0 -
For what it is worth, I don't really care with A-Levels - I knew mine were much easier at the time. I object to the drop in degree standards as mine wasn't easy.0
-
Possibly, but a risky assumption. You’d need to consider candidate numbers and other socio-economic factors to be sure.TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%
Maybe a greater number of capable students were choosing to study than had previously.
UK in 1979 was a very different place to UK in 2000, just ask Rick what influenced it all.
0 -
The nature of the exams changed. The 80s were norm referenced and GCSE was designed to be criterion referenced so as teachers became more familiar with the exams, marks went up and if students jumped through the hoops they got the grade (broadly speaking).TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%
Modular exams helped a lot at A-level as did the ability to retake modules as often as wanted. What did change was the amount in the course. By taking away most of the summer term to do AS levels, there had to be a reduction in the amount of material covered in the specifications. Hence exams were “easier” because there was less in them.
The advantage of the IB is that there are no year 12 exams so the amount of material remains challenging. Higher maths is a bit more like the old A-level in terms of what’s covered but even there old A-level maths had more in it.
I agree about degree courses. I’m pretty sure that the grade inflation there is driven by a sense of entitlement.
BTW, I did O-levels in 1985 and did well for the time, A-levels in 1987 and got A in maths, which meant something then (!), D in Physics, which was average at the time, and N in Chemistry (for nearly passed rather than failed!). My degree came from an old university at a time where there were universities and polytechnics and I got the average class for the time (a Desmond). These days, average A-level grade is B and average degree class is a high 2:1. Students these days are not brighter than I was, honest!0 -
This is just correlation and could equally be explained by any number of things including *better teaching*TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%0 -
Assuming the structure of the exam hasn’t changed then the papers will be of the same difficulty from year to year.Dorset_Boy said:
That might be the objective but in reality it's not the case, hence adjustments at the grade boundaries each year under the current system. Additionally someone earlier suggested that exams from a number of years ago were of a different standard.rick_chasey said:Well the exams are supposed to be the same level of difficulty between the years so an A in 2015 is as difficult as an A in 2019
If you were to go for a grading system which was entirely to do with distribution how would you measure improvement in students grades at schools? It would be entirely zero sum which helps no one.0 -
All an employer or University can ever do is look to pick the best of each year group though. Knowing that this year's A level group are better or worse than their counterparts 5 years ago isn't much use if you're offering a place in HE or on an apprenticeship. You could even do away with grades and simply allocate the person's percentile ranking. If you did need to assess against previous years you could presumably look at what percentage was required to get the same grade in previous years.rjsterry said:
Is that not just potentially hiding any changes in the standards of exams? The results of each cohort would look exactly the same regardless of whether the standard of teaching or the difficulty of the exam went up or down.Dorset_Boy said:There are now 9 grades at GCSE. Why do 11% of pupils not fall into each grade?
Likewise, if there are A-E + F grades at A-level, why aren't 1/6th of student in each grade?
Surely the point is to see where the individual lies in relation to their peer group and this means no grade inflation, just an easier assessment for those looking in from the outside as to where a person with say grades from 2015 compares with someone with grades from 2019.0 -
I think the result of all this is that final exams will be phased out and the system will move to a more continual assessment process and interim testing which will give a better basis for grade estimation if required in the future.
It would never have suited me as I was one of those lazy people who didn't bother working but could memorise enough to get decent grades in a final exam. However, it seems a far better way of doing things for numerous reasons. It takes the pressure off so much resting on a single exam, it gives a far better basis on which Universities can assess candidates rather than waiting until mid August to make a final offer for a place, it rewards students for consistent performance rather than my old approach and it's more in line with the way HE courses are assessed.1 -
I do wonder how technology will start to influence things.
Centralised electronic multiple choice assessments maybe!
For some reason, there is a political distrust of the teaching profession which has manifested itself in wholesale meddling beyond just budget management as in the NHS.
I can’t forgive Gove for his misguided assault on education. The man is such a dangerous idiot.0 -
Here's some academic research on it. Consistent standard over the last 20 years, but higher stanard before that.rick_chasey said:
This is just correlation and could equally be explained by any number of things including *better teaching*TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%
https://bera-journals.onlinelibrary.wiley.com/doi/abs/10.1002/berj.32240 -
That being said, what is the material impact? I know my brother found that the easier maths A level course meant he found his first year of Mechanical engineering at Imperial a bit heavy going but once he caught up it was a non issue.TheBigBean said:
Here's some academic research on it. Consistent standard over the last 20 years, but higher stanard before that.rick_chasey said:
This is just correlation and could equally be explained by any number of things including *better teaching*TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%
https://bera-journals.onlinelibrary.wiley.com/doi/abs/10.1002/berj.32241985 Mercian King of Mercia - work in progress (Hah! Who am I kidding?)
Pinnacle Monzonite
Part of the anti-growth coalition0 -
Spoken like a true blue tory stooge.coopster_the_1st said:
How would your model cope with the 25 schools where every teacher predicted every student an A* in every single subject?morstar said:My bugbear is this.
I have built quite a few data models over the years and completely accept that any assumption based model is going to have some flaws.
However, when you run your model and identify the deviations in some students and institutions that are huge, it is obvious there is a fundamental lack of finesse in your model so you refine it.
This is not massively complex data modelling and you have had the data for months. It is simply inexcusable for it to be so poor.
Grade inflation. The ongoing debate.
Are kids brighter today? Not really.
Is teaching better today? On the whole I think aspects of it are.
Are schools more effective at achieving exam results. 100% yes as that is how they are measured.
Does this make a better education. Not necessarily, it makes a narrowly focussed education that is particularly ill suited to those with less academic abilities.
Apart from the stupid panic decision to cancel the exams, the problem is the teaching profession and their lack of professionalism, and it is the student that gets hurt by their unprofessionalism.1 -
There are many academies in particular that will have over cooked their predictions for pupils because it has been just another irresistible potential opportunity for them to game league tables and headlines for their marketing and careers of those in leadership of these schools.
Academy heads all too often already game and have institutionalised the gaming of league tables through off-rolling, whether that be talking lower performing kids out of continuing with A-Level courses that they aren't doing so well in, or talking parents into moving their kids to another school.
As soon as the exams were called off, it was always going to happen that many academies would overcook their predictions under the pressure of questionable management. It must have been incredibly uncomfortable for the teachers at such academies.
The whole system was always going to have attempts to game it, so even the data going in would be questionable in many cases and that's before all the ridiculous ways the 'algorithms' have been evidenced to work so poorly and inconsistently.0 -
Good question.rjsterry said:
That being said, what is the material impact? I know my brother found that the easier maths A level course meant he found his first year of Mechanical engineering at Imperial a bit heavy going but once he caught up it was a non issue.TheBigBean said:
Here's some academic research on it. Consistent standard over the last 20 years, but higher stanard before that.rick_chasey said:
This is just correlation and could equally be explained by any number of things including *better teaching*TheBigBean said:A quick google seems to support the idea that standards fell a lot between the 80s and the 90s.
Do do more people get an A today at A level than they used to?
Yes, according to stats published in The Spectator this week which show how the proportion of candidates achieving A (and A* since 2010) grades has increased over the years:
1982 8.9%
1994 18%
1999 24%
2004 28%
2009 31%
2014 (A & A* grades) 26%
2019 (A & A* grades) 26%
https://bera-journals.onlinelibrary.wiley.com/doi/abs/10.1002/berj.3224
I think if standards have been constant for 20 years then it is less of an issue, but note that the same thing is happening to degrees now.
I think there are three downsides from this. Firstly it makes comparing different individuals of different ages harder although you could probably argue that there is a sufficient time gap in the deterioration for this to be pointless anyway. Secondly, it removes the ability of a bright student to demonstrate their skills. Finally, it's surely better to learn more than less.
On the last point, it is worth noting that some studies have shown Korean children to be a full academic year ahead of their British peers. Now, there's a lot more to that and the Korean system has many downsides, but it's still not that great for the UK.
As I have said before, my A-Levels were of the easy type, so I'm not just saying that the kids have it easy today.0
