The way forward for LC2021

Bobtheman · 25-03-2021 02:52PM

The departments gross incompetence is actually reassuring to me. Looking on them as your enemy which they have become since being captured by non teaching academics and managerialism - it means they find would find it difficult to really **** up teaching by forcing through mandatory paper work. The underfunding plays a part to.
We are a long way away from paid subject department heads.
Their heads up their arses is shown by the document we just got today. More stuff on accredited grades.
There are paragraphs where they encourage you to ask higher order questions with wait time blah blah
There is not one line of useful information in the section on English. I'd say its the same elsewhere.

Wombatman · 25-03-2021 02:52PM

You are making some horrible assumptions here.

Purefrank128 wrote: »

FALSE!

There is no reason to think that the cohort taking the exams will have the same ability as the cohort taking accredited grades.

There is no reason the think otherwise either. Students who sit exams should not \ will not be given preferential treatment.

Everyone - or almost - everyone is going to get an accredited grade. I think it's safe to assume that most students who sit the exams are at the top end - they think their exam result could improve their grade and they don't mind doing the work to show that.

This may be borne out in exam results. It wasn't really last year. Either way it doesn't matter to the process.

The grade distribution of the group sitting exams should be better than that of the group getting an accredited grade (i.e. everyone), which should be better than that of the group getting an accredited grade only.

It may but both should be standardised accordingly and on their own.

But none of these grade distributions should be like those produced by exams in 2019 and previously.

Interesting one here. You have three grades.
Accredited grade, exam grade and final grade (the best of both).

Should the final grades be further standardised to fit the curve? I guess it would have to be.

Bobtheman · 25-03-2021 03:42PM

All grades will be standardised to fit the curve particularly at honours level.
Predicted grades were put through algorithms and Bell curves last year. It will happen this year.
Even those who sit a physical exam in June will be filtered through a bell curve. I'm sure some schools have better years etc but overall the bell curve is the bell curve
Whatever is the best grade = final grade. Filters used by that stage.
There was a lot of bull**** in the press that Deis schools were not adjusted downwards at all last year.
It did happen. But it seems middle class schools were harder hit.

Random sample · 25-03-2021 05:48PM

https://assets.gov.ie/128197/f98845d8-4ced-4b51-bea7-ac544f41de88.pdf

Is anyone else sick of reading long documents?

am_zarathustra · 25-03-2021 07:22PM

Wombatman wrote: »

They force it every year. That's the whole point of standardisation.

Take the Leaving Cert exam in Higher level English as an example.

The proportion of students who achieved a H1 (90 to 100 per cent) in 2019 was 2.9 per cent; the equivalent figure for 2018 was 2.9 per cent; in 2017 it was 2.9 per cent.

The pattern is repeated further down the grades.

The proportion of students who achieved a H2 (80 to 90 per cent) in 2019 was 10 per cent; in 2018 it was 10 per cent; and in 2017 it was 10.7 per cent.

Similarly, the proportion of students who achieved a H3 (70 to 80 per cent) in 2019 was 20.4 per cent; in 2018 it was 20.6 per cent.

Normally they standardise using adjusted making schemes relative to emerging trends that year. How they do it this year is anybody's guess.

This is doable in english, there's a subjective element. The number of H1s vary year on year in sciences, one particulatly doable year with a very direct questions led to almost 12% getting H1s, following year it was less than 8%. If the answer is lumen, it's lumen.....I can weight all I want but good students will ace the biology exam bar the questions are much harder than normal. It wouldn't be uncommen for kids to get 60/60 on a couple of questions normally.

Even maths is more workable as there are always really hard bit that aren't exactly preparable for, these will be used to sort the H1s-H3s

Treppen · 25-03-2021 08:14PM

Leftwaffe wrote: »

There is a line in the 'SEC Oral guidance to school management' which states the following:

"If candidates want to present for interview and who are not listed on the roll, the examination aide must write their name, examination number and date of birth on the blank page of the roll"

Would this suggest they can still show up and do the oral despite opting for an accredited grade?

I always thought that was for externs who couldn't find a centre. I had that issue before for extern music students when a school pulled day before. They found another school last minute and the exams aid was instructed to just add in their name.

Treppen · 25-03-2021 08:22PM

Purefrank128 wrote: »

FALSE!

There is no reason to think that the cohort taking the exams will have the same ability as the cohort taking accredited grades.

Everyone - or almost - everyone is going to get an accredited grade. I think it's safe to assume that most students who sit the exams are at the top end - they think their exam result could improve their grade and they don't mind doing the work to show that.

The grade distribution of the group sitting exams should be better than that of the group getting an accredited grade (i.e. everyone), which should be better than that of the group getting an accredited grade only.

But none of these grade distributions should be like those produced by exams in 2019 and previously.

Ok I'll have to call in the maths and stats gang but from my understanding once you've about 20,000 thousand students in two different camps then ability is very similar. I doubt it'll be all the brainboxes Sitting exams.

Treppen · 25-03-2021 08:27PM

Bobtheman wrote: »

All grades will be standardised to fit the curve particularly at honours level.
Predicted grades were put through algorithms and Bell curves last year. It will happen this year.
Even those who sit a physical exam in June will be filtered through a bell curve. I'm sure some schools have better years etc but overall the bell curve is the bell curve
Whatever is the best grade = final grade. Filters used by that stage.
There was a lot of bull**** in the press that Deis schools were not adjusted downwards at all last year.
It did happen. But it seems middle class schools were harder hit.

your forgetting about extenuating circumstances.
As I said before you can't get the percentage of H1s down to fit a standard bell curve if it involves making questions incorrect when the student got them correct.

am_zarathustra · 25-03-2021 08:39PM

Treppen wrote: »

Ok I'll have to call in the maths and stats gang but from my understanding once you've about 20,000 thousand students in two different camps then ability is very similar. I doubt it'll be all the brainboxes Sitting exams.

There will be an ascertainment bias on the exam cohort, you train the set sitting the exams (their AG) on the overall AG, so you bell curve the exams to the AG roughly but only the cohort of both, the intersection. You'd have three data sets, one of all the candidates AGs (group A), one of the candidates sitting the exams (group

and then the exam results from the exams (group C).

Initially you get the national standardized curve for group A, then the curve for group B. You compare, are they actually different (literally like the HL stats Ho testing but obviously with better, more nuanced stats but honestly a standard z test would nearly tell you anyway). If there is a significant difference (again, we can't assume this but experience would tell me it'll skew towards the top) then this is the curve you aim for with exams, or something similar. So group A and B compared, group C then normalized off B not A provided their is a significant difference. So the national curve has H4 as the mean (say 55% or something), group B has a mean of 67% from AG, a reasonable assumption is that group B has more high achieving students and should therefore have a skewed curve when corrected. Now spread will be another thing to deal with but you'd hope that would take care of itself.

This is how I would see it running I think, but I haven't thought it out fully obviously, it's be nice to know though.

rainbowtrout · 25-03-2021 09:05PM

Purefrank128 wrote: »

FALSE!

There is no reason to think that the cohort taking the exams will have the same ability as the cohort taking accredited grades.

Everyone - or almost - everyone is going to get an accredited grade. I think it's safe to assume that most students who sit the exams are at the top end - they think their exam result could improve their grade and they don't mind doing the work to show that.

The grade distribution of the group sitting exams should be better than that of the group getting an accredited grade (i.e. everyone), which should be better than that of the group getting an accredited grade only.

But none of these grade distributions should be like those produced by exams in 2019 and previously.

That would be reflected in the numbers that sat the winter exams. All the four sciences were in the most popular subjects in terms of numbers that sat the exam. More students sat Chemistry than English. I'd imagine a lot of those were the Medicine/Dentistry/Veterinary crew looking for H1s.

page 8 shows the numbers that sat the winter LC for each subject.

https://www.examinations.ie/misc-doc/BI-PR-78526925.pdf

rainbowtrout · 25-03-2021 09:14PM

Wombatman wrote: »

True. The distribution of marks should follow the same pattern of previous years. This could involve big shifts based on only a couple of marks because of the narrow test base this year.

Simple example:
If 20% of sitting students of are given got a H3 in years gone by, give or take 1-2%, then for AG marks and exam marks, 20% of students should get a H3 give or take a couple of %.

This could still result in an individual student a getting an exam grade of H4 and an AG grade of H2 but that's a different story.

It won't happen like that. Currently I have two Leaving Cert groups for different subjects. In one class 80% of them are opting out of the exam, in the other 80% are opting in to the exam. The 20% that are opting in to the exam in the first class are the best academically and are keen to secure the top grade in case the grades come down through standardisation. The 80% opting out are largely not performing above a mark of 80% in class tests, so know they are relatively safe one way or the other. And quite a few of them if offered a H6 in the morning would take it and be out the school gate in a flash.

In my physics class yesterday I gave them one of their monthly assessments, most are high achievers and competitive with each other and after the test they immediately started a post mortem on the questions on the test. All of them will sit the exam because they know grading will be tight and even with me grading them fairly, they know I can't guarantee that grades won't be adjusted so will sit the exam to prove their ability.

Wombatman · 25-03-2021 10:52PM

rainbowtrout wrote: »

All of them will sit the exam because they know grading will be tight and even with me grading them fairly, they know I can't guarantee that grades won't be adjusted so will sit the exam to prove their ability.

Exam results will have to be standardised too though. They have to be.

Take a fictional scenario where the range of marks in the Physics HL exam is between 100% and 70%. No student is scoring outside of this range. This is using default marking scheme A.

If traditionally 3% of all students get a H1, for example, and in our high performing group, 3% got between 100% and 98%, then the marking scheme A will be refined to marking scheme Z, to ensure that 100% to 98% becomes 100% to 91%.

If traditionally 10% get a H2 and the next 10% are between 97 and 92 under marking scheme A, then marking scheme Z will shift this group to between 90% and 81% and so on.

I'm not sure how it will be achieved systematically but it will have to follow these lines or college points will see chronic adjustments which will mitigate against the 2020 students, who deferred for a year.

Their will be major standardisation casualties because of it's bluntness.

I expect the final "best of both" result will also have to be standardised too as it will skew to the top.

Say these are the final normalised results for a subject coming out of both independent processes. (Process have to be independent because some students wont do the exam and some won't get AGs (home schoolers).)

Student - AGs - Exam Grade
S1 - H1 - H2
S2 - H2 - H3
S3 - H3 - H1
S4 - H4 - H5
S5 - H5 - H4

Final Grade
S1 - H1
S2 - H2
S3 - H1
S4 - H4
S5 - H4

So here, AGs are nicely spread and exam grades are nicely spread, but the final mark is skewed to higher grades.

am_zarathustra · 25-03-2021 11:11PM

Wombatman wrote: »

Exam results will have to be standardised too though. They have to be.

Take a fictional scenario where the range of marks in the Physics HL exam is between 100% and 70%. No student is scoring outside of this range. This is using default marking scheme A.

If traditionally 3% of all students get a H1, for example, and in our high performing group, 3% got between 100% and 98%, then the marking scheme A will be refined to marking scheme Z, to ensure that 100% to 98% becomes 100% to 91%.

If traditionally 10% get a H2 and the next 10% are between 97 and 92 under marking scheme A, then marking scheme Z will shift this group to between 90% and 81% and so on.

I'm not sure how it will be achieved systematically but it will have to follow these lines or college points will see chronic adjustments which will mitigate against the 2020 students, who deferred for a year.

Their will be major standardisation casualties because of it's bluntness.

I expect the final "best of both" result will also have to be standardised too as it will skew to the top.

Say these are the final normalised results for a subject coming out of both independent processes. (Process have to be independent because some students wont do the exam and some won't get AGs (home schoolers).)

Student - AGs - Exam Grade
S1 - H1 - H2
S2 - H2 - H3
S3 - H3 - H1
S4 - H4 - H5
S5 - H5 - H4

Final Grade
S1 - H1
S2 - H2
S3 - H1
S4 - H4
S5 - H4

So here, AGs are nicely spread and exam grades are nicely spread, but the final mark is skewed to higher grades.

It would still make sense to train your set on the AGs to get your curve for each exam. It would lead to fewer discrepancies if it's weighted up and reduce the bluntness quite a bit. It's would only require the mean and SD to work off, the SEC chief and advising examiners would be well used to it, you have the normal curve. They are less interested in H1s in most subjects too, the changes to schemes generally focus very much in the middle, sort the center and the rest sorts itself. You standardize from the center

deiseindublin · 25-03-2021 11:38PM

I can't see how the AG and Exam results can be married, with a much smaller and different cohort sitting the paper. As a previous poster said, it will probably be the high achievers that choose to sit the exams, the H1s and H2. I just can't see a lot of H5 and H6 candidates bothering, unless they're so borderline H7 that they feel they have to.

How do you use a bell curve without a complete set of data/cohort of students, and some grades missing in their entirety?

am_zarathustra · 25-03-2021 11:46PM

deiseindublin wrote: »

I can't see how the AG and Exam results can be married, with a much smaller and different cohort sitting the paper. As a previous poster said, it will probably be the high achievers that choose to sit the exams, the H1s and H2. I just can't see a lot of H5 and H6 candidates bothering, unless they're so borderline H7 that they feel they have to.

How do you use a bell curve without a complete set of data/cohort of students, and some grades missing in their entirety?

It's would be a standardized curve with a specific mean and a specific SD, that's actually straightforward. You take the cohort that are sitting their exams, take their accredited grades as a whole and get the mean and SD, use that as your normal curve. Then try and get close enough with the marking schemes alterations.

The real issue will come if the AG are insanely overinflated, they will need to be standardized before you remove the exam sitting cohort to get their "predicted curve". You'll still end up with more higher grades, but if you give two options I just can't see how that will be avoided without throwing the dataset toget there at the end and standardizing again, which you cant do coz the exams will have a grade on the front that can't be changed (digitally or literally)

Without knowing it's safer to do both anyway

rainbowtrout · 25-03-2021 11:59PM

Wombatman wrote: »

Exam results will have to be standardised too though. They have to be.

You can't standardise in a normal way whether you follow the traditional bell curve (pre 2020) or the inflated one from last year. The data set is too small.

The H1s and to a lesser extent the H2s are where the inflation in the predicted grades. Having said that. If a student is a genuine H1, then the teacher will have given that student that grade. If they sit the exam and achieve it they have proved their ability.

am_zarathustra · 26-03-2021 12:17AM

Bar less than 100 students are sitting an exam you can standardize, why couldn't you?

There's the normalize versus stadardize argument for sure but you'd need to see the first AG dataset to really make that choice. I would assume they know what the normal distribution of grades in any subject generally is anyway. Very few subjects will have small enough numbers to stop standard stastics working just fine, maybe the odd niche OL paper but they won't impact inflation all that much anyway. Grade inflation will really only matter in 15/16 subjects based on the numbers taking them.

DubLad69 · 26-03-2021 12:55AM

Somewhat off topic here, but does anyone know how many days large schools can usually hire an examinations aide for?

Just the normal exams aide, not the new one for the orals or the calculated grades aide.

rainbowtrout · 26-03-2021 01:21AM

DubLad69 wrote: »

Somewhat off topic here, but does anyone know how many days large schools can usually hire an examinations aide for?

Just the normal exams aide, not the new one for the orals or the calculated grades aide.

Not 100% sure but I think it’s about 15. They have to be about for every day of the exams and the day before they start for set up I think

Leftwaffe · 26-03-2021 10:42AM

The SEC have told me that students can do the oral who have opted for the predicted grade. So anyone can do the oral essentially. They're just added to the end of the examinations aide roll.

Random sample · 26-03-2021 10:57AM

Leftwaffe wrote: »

The SEC have told me that students can do the oral who have opted for the predicted grade. So anyone can do the oral essentially. They're just added to the end of the examinations aide roll.

Assuming you have an interviewer available? Not every school would have an interviewer available for longer than they initially signed up for.

Bobtheman · 26-03-2021 08:15PM

I'm rarely proud OF the Asti but was proud they told the government to **** off in January
Having agreed to assess our own pupils hand over fist last year I really hope we ain't setting a precedent

Str8outtaWuhan · 26-03-2021 08:19PM

Bobtheman wrote: »

I'm rarely proud OF the Asti but was proud they told the government to **** off in January
Having agreed to assess our own pupils hand over fist last year I really hope we ain't setting a precedent

Last year asti members agreed to provide an provisional grade that was not certified by the SEC.

This year asti members re working over easter, working saturdays for lcvp/comp sci exams and assessing their students for an SEC accredited grade. do the math

Bobtheman · 26-03-2021 11:13PM

Str8outtaWuhan wrote: »

Last year asti members agreed to provide an provisional grade that was not certified by the SEC.

This year asti members re working over easter, working saturdays for lcvp/comp sci exams and assessing their students for an SEC accredited grade. do the math

I am still proud of the Asti regarding the events in January'. Given the bleeding heart narrative over kids it was hard to avoid assessing our own pupils. I'm just hoping we are not out maneuvered long term on this issue. We are a long way from the UK bull**** in schools. To pretend we ain't is to be delusional.
A global pandemic is a global pandemic. Its up to union members to say no to allowing the department to use this crisis to effect long term change. But you could imagine the outcry from nurses if teachers refused any of the above days.

Leftwaffe · 26-03-2021 11:22PM

For anyone who’s doing the oral aide job, has there been payment forms released yet? Haven’t seen them anywhere.

DubLad69 · 27-03-2021 10:27PM

Leftwaffe wrote: »

For anyone who’s doing the oral aide job, has there been payment forms released yet? Haven’t seen them anywhere.

Your school should pay you directly and claim it back from the SEC.

Leftwaffe · 29-03-2021 08:20PM

I have approx. 120 students on my roll for the orals but only about 75ish are actually doing it and timetabled to do it. Why is there this many on my roll and what do I write beside the name of those who aren’t doing it? Did not attend?

Enright · 29-03-2021 09:10PM

"Withdrawn" for those that pull out "Did not attend" for those that do not turn up

Leftwaffe · 29-03-2021 09:55PM

Enright wrote: »

"Withdrawn" for those that pull out "Did not attend" for those that do not turn up

Thanks. I’ve had no withdrawals and all students on my roll were made aware of the orals and given an option to be timetabled. I assume “Did not attend” is my option here then.

Enright · 30-03-2021 01:07PM

The only students on the official roll we be ones that engaged with the online system and opted in AND students enrolled in your school on P-Pod and who did not engage to withdraw themselves, note on the roll instructed the school to make contact with these students to determine if they are going ahead with it or not. Principal/deputy principal to sign off on the quality of each recording.

NB: Students not on the roll are entitled to turn up on the day and you are obliged to add them to the roll.

The way forward for LC2021

Comments