2020 A Level Results and moderation

[ClarasZoo]

My son is awaiting the result of one A level this year. He is predicted an A and has had A exams/assessments throughout the year. Following on from the thread on the Scottish Highers disasters I am starting to worry about the results, both for him and for two family friends in the year above.
Can anyone who understands the system better than me please advise/reassure me?
So, the first example is a child who needs an A grade in one particular subject to go to the uni they want. That child is predicted an A/B. there are 5 pupils doing that subject at his school. Last year one got a B, two got Cs and two got Ds. the two years before were similar. This year, for some reason, it is a clever cohort of 6. 5 are predicted A stars and one A(friend's son). Based off previous years, there is only one B, 2 Cs and 2 Ds to hand out - can that be right? In that case, my friend's son is going to get a D, even though he should get an A/B. Please tell me this is not how it works, although looking at the press it does seem to be.
My other example is a girl who needs an A star in Italian A level. She is Italian so would breeze an A star. However, her school has never got more than a C in Italian. So will she get a C? That CAN"T be right?!
My son's grade is clearly also in doubt, if the school does not have sufficient to "hand out", isn't it?

[titchy]

That's not how it works!

There's been a million threads on this already which have explained the principles. We also now know 96% of actual grades awarded will be within one grade of all teacher assessed grades.

[titchy]

Btw the Italian girl who needs an Astar in Italian should check her uni entry requirements. It's very unusual for a uni to accept language A levels from a native speaker.

[deFleury]

there are 5 pupils doing that subject at his school.
He should be fine then. Small cohorts are protected as there isn't enough data for the algorithm, with "very small" cohorts getting their CAG with no moderation at all, and then weighting on a "sliding scale" between CAG and algorithm as they get a bit larger. None of the exact numbers involved have been specified, but 5 must be pretty small, especially if there are similarly low numbers in prior years.

[deFleury]

As for your son, assuming a regular size cohort (I've looked for data but can't work out what this is) - if his teacher has overestimated A compared to the prior three years then unless that subject cohort also got significantly higher GCSE results than prior cohorts, these will be moderated down. In that case what will make most difference to your son is his position in the A ranking his teacher made. The higher he was listed, the less likely he is to be moderated downward.

[ClarasZoo]

Thank you for some reassurance on the small cohort! The Italian eg was so as not to identify. The girl is not doing the language at uni but still needs the A star... but has other A levels too. My son is not a small cohort unfortunately but I have no idea whether they will rank the One year A level takers with the two year and so he has no idea of his position...

[ClarasZoo]

Titchy- please can you point me to any basic info then? I couldn’t find that many threads actually.. has that 96% figure been published? Because I also have a daughter awaiting GCSEs... would like to know if also 96% for that...

[ClarasZoo]

Actually I have just read that for subjects over 15 entries this is exactly how it works- CAGs are being ignored... it’s all just historic data and ranking. Under 5 entries will be CAG. Between 5-15 it’s a bit of both...

[ClarasZoo]

It’s published on TES today. So in my silly example if 16 people are taking Italian but the school has never got more than a C at Italian, the Italian fluent speaker would get a C. This model will favour private schools with small subject cohorts- it’s not fair..

[augustsong]

Are they really completely ignoring CAGs for cohorts if over 15?

Surely though, “historic data” doesn’t only mean the past A-level results of a given school. Would it nor also mean that they would look at the GCSEs of those candidates? If they were higher than past cohorts, then that is evidence that the A-levels would be too?

[ClarasZoo]

If you look at TES today yes they seem to be ignoring everything except historic data...

[ClarasZoo]

Apologies- they will look at historic attainment of cohort. But in my Italian example, that girl will get a C. If she was at a private school with a three kid Italian entry she would get an A star. How is this fair?!

[JacobReesMogadishu]

I saw a tiktok video from a girl in this position......ok not the best source I know.

She got all As and A IN her GCSEs and in her A level mocks. Has also scored high in every in class test. She needs As and A for her uni course.

She goes to a very under achieving school with poor grades and her teachers have told her to expect Cs, maybe a B or 2 if she's lucky. Now obviously the teachers may be wrong and I hope to God they are because this seems very unfair. She said exactly the same - if she went to a better school she would get the As.

[Frazzled6]

I understand that : Subject cohorts of under 5 will most likely be awarded their CAGs. Cohorts of 5-15 will have the CAGs taken into account. Cohorts of 16 will have the results based on historical data. There will be some exceptions.

[augustsong]

But that sounds ridiculous. Surely, in the case of the Italian girl OP knows, she would also have got a 9 at GCSE in Italian? That is far more solid evidence of her likely grade at A-level than what other random students may or may not have achieved in past years?

[deFleury]

There will be some exceptions.

I haven't seen anything about exceptions. Do you have a source?

[deFleury]

Surely, in the case of the Italian girl OP knows, she would also have got a 9 at GCSE in Italian? That is far more solid evidence of her likely grade at A-level than what other random students may or may not have achieved in past years?
I think I've ready about one regulator (Welsh?) taking an approach including the individual student GCSE results, for A level. OFQUAL decided not to. They have taken the mean GCSE grade (across all subjects) for the whole group entered for the subject at that school, and compared it to the mean GCSE grade for whole group who got A levels result in that subject and school over three years. So if this year's group got a mean of 6.4 and last years was 5.2 then A levels will be expected to be somewhat higher than last years.

[augustsong]

Also, it’s one thing to use historical data for selective grammar or independent schools, but results at non-selective state schools must vary massively year to year?

For instance, my DC is at a school that gets between 70-75% A*-A every year. It’s probably quite easy for the school to rank students within subjects and identify the ones who would be on track for the B grades. Such a school would have the consistent historical data stretching back decades. But for a school with the full spectrum of ability range, results must vary year to year depending on who happened to be in the catchment area at the start of year 7?

[Coffeeandbeans]

That is exactly how I understand it. My child’s schools A level results generally are not great. However they are improving and have always said this years cohalt are bright. This means nothing on the day and for my child’s economics and physics it will be based on historical data. So very very unfair.

[ClarasZoo]

Right so I am now very concerned about my daughter's GCSE results too. In my Italian friend's case, it does not matter that she is excellent at Italian (obviously) because historically her school does badly and it's not a bright year. So she will get a C, basically. I suppose her remedy is to do it in October and get an A* but this will mean taking two years out (she is a medic and needs it). In my daughter's school (private) they got 8/9s for Chemistry (triple) last year. So even the kid who is predicted a 5 and has messed about all year will get an 8. How is this system not benefitting private schools with small cohorts? It must be more discriminatory towards children in state schools with large cohorts? How can OFQUAL justify this? OFQUAL - anyone reading this?! Can individual schools or pupils appeal?

[SeasonFinale]

The figure of 96% being within one grade relates to the Scottish results. The TES article referred to above suggest that there is 10% over estimation (unlike to 20-25% in Scotland) and that most of these are at pass/fail level for gcses presumably where a teacher will not want to give a 3 to someone and has gone for a 4 and at a grade B level for A levels.

Please note the Scottish results will have absolutely no bearing at all on how the English awarding boards will be doing their standardisation.

Also as long as the school has sensibly assessed their other grades within the basic historic and prior cohort data they will allow an outlier grade for a high performer. Thus a single or couple of very high achievers should still get their high grade. The issue would come if the school decides to act in a daft manner and tries to submit higher grades across the board!

In larger cohorts previous historic results shows there is very little difference from year to year even when exams are being taken.

Further the part that has been added to the Appeals today in the JCQ Guide to appeals published today on their website shows that if the school can show a significant difference to why a specific cohort or group should get better grades then they can appeal on this basis.

People also need to remember that UCAS predicitons are generally very best case scenario grades and are often inflated as aspirational grades rather than actually achievable grades so look at working at grades or expected grades as a truer reflection of what they should expect.

People need to just wait and see what they get, they may be presently surprised.

[SeasonFinale]

Please also remember that Prior Cohort attainment is an additional layer as well as national attainment (and this would have given higher grades anyway at a national level).

[deFleury]

Also as long as the school has sensibly assessed their other grades within the basic historic and prior cohort data they will allow an outlier grade for a high performer. Thus a single or couple of very high achievers should still get their high grade. The issue would come if the school decides to act in a daft manner and tries to submit higher grades across the board!

Where have you seen this in the published information from OFQUAL, please? Allowance for outliers had been guessed, but I can find no mention of it anywhere. In fact, quite the opposite.

[mumsneedwine]

www.gov.uk/government/news/response-to-tes-story-about-centre-assessment-grades

This is helpful

[MillicentMartha]

The blog linked also says about schools being allowed to appeal for outliers if the moderation doesn’t work for them.

ofqual.blog.gov.uk/2020/08/06/fairness-in-awarding/

[ClarasZoo]

@SeasonFinale

The figure of 96% being within one grade relates to the Scottish results. The TES article referred to above suggest that there is 10% over estimation (unlike to 20-25% in Scotland) and that most of these are at pass/fail level for gcses presumably where a teacher will not want to give a 3 to someone and has gone for a 4 and at a grade B level for A levels.

Please note the Scottish results will have absolutely no bearing at all on how the English awarding boards will be doing their standardisation.

Also as long as the school has sensibly assessed their other grades within the basic historic and prior cohort data they will allow an outlier grade for a high performer. Thus a single or couple of very high achievers should still get their high grade. The issue would come if the school decides to act in a daft manner and tries to submit higher grades across the board!

In larger cohorts previous historic results shows there is very little difference from year to year even when exams are being taken.

Further the part that has been added to the Appeals today in the JCQ Guide to appeals published today on their website shows that if the school can show a significant difference to why a specific cohort or group should get better grades then they can appeal on this basis.

People also need to remember that UCAS predicitons are generally very best case scenario grades and are often inflated as aspirational grades rather than actually achievable grades so look at working at grades or expected grades as a truer reflection of what they should expect.

People need to just wait and see what they get, they may be presently surprised.

This is very helpful - thank you. If there has been 10% over estimation of grades does this mean you only have to worry if you are number 10, say, of grade 6s awarded? Meaning you are then the one to get bumped down to a 5? The other 9 x 6s are "safe"?