45% of students set to receive an 'unreliable' English GCSE grade on Thursday

[noblegiraffe]

Well this is a worrying story.
Basically, due to marking inconsistencies, especially in a subjective subject like English, grades which fall close to a grade boundary are unreliable, because a different marker could have legitimately marked it differently.

The majority of students used to receive a grade A*-D, so 4 grade boundaries. Now that will be from 9-3, so 6 grade boundaries. As the number of grade boundaries increases, the number of students who fall close to a grade boundary and thus have an unreliable grade increases. This represents an increase from 30% of grades being unreliable to 45%.

The article suggests that grades should be scrapped, as the difference between a 79 and an 80 is slight, but one mark could get a B and the other an A, which will be viewed differently.

www.telegraph.co.uk/education/2017/08/20/thousands-students-set-receive-wrong-gcse-mark-new-system-experts/?utm_source=dlvr.it&utm_medium=twitter

Obviously the same problem will affect maths, but the marking is usually more objective.

[Notanothergiraffe]

Oh blimey DS is with AQA. so the schools will all know the marks but they won't tell he pupils automatically. I think that is shit, at the end of the day who's results are they?!

[tararabumdeay]

Percentage scores would be much fairer. Of course we don't know grade boundaries for this year but in previous years a student with 60% of available UMS marks would get the same grade as one with 69.7%.

The grade boundaries 'lower, middle, upper' could be published each year with the percentage range alongside.

A grade 2 to a grade 4 pass could mean up to 90 UMS marks or 30% improvement in old A - G grade boundaries. That's a huge jump now that Ofsted are looking for individual progress but centres will still be judged on pass/fail.

Centres will have % grade awards available on Wednesday.

This pass/fail distinction is so unfair as GCSE was originally introduced to be one exam for all.

[Witchend]

I have brief daydreams where they all get 5s (hah, no chance, some of them had a quite challenging Y11) and I'm carried into the first INSET day on the shoulders of SLT, but then the shadow of them all failing and me being fired for extreme incompetence takes over.

That made me laugh. I have similar two extreme case scenarios going round my head for dd1's results.
I'm nervous they're going to make a point and mark them all harshly. Then slacken off afterwards leaving this year with what looks like really poor results with the same levels.

[Cafeconleche]

I read the Telegraph article and felt sick. My DS needs a 5 in maths and 6 in English Language (plus Bs in 4 other subjects) to get into his 6th form. He has achieved this (and much higher) in all his mocks, but the thought that 45% of the grades could now be wrong is not only distressing, it's a bloody outrage. I hope there is a very special place in Hell reserved for Gove. I know how hard my DS has worked and what he deserves for his efforts - and also how much time and effort his teachers put in, which was above and beyond. I keep saying to myself that this time next week it will all be over and my DS will be where he is supposed to be. But unfortunately it's not a level playing field. What an utter fucking mess.

[woman12345]

People who study GCSE-itis draw a distinction between “criteria” referenced and “norm” referenced exams. Criteria testing fixes what is to be tested before the test is sat, and whoever achieves the criteria passes. The people responsible for GCSE-itis – exam boards, overseers of exam boards and people such as you, Ms Greening, who oversee the overseers – have come up with a better system. First they test, then they look at how the testees have performed. If it looks as if too many or too few have done well, then they move the grade boundaries. In other words, they raise or lower the high-jump bar; they fix the “norm

www.theguardian.com/education/2017/aug/22/gcse-grade-boundaries-altered-michael-rosen


“For independent schools, that’s going to be a difficult message for them to have to explain to parents. There have been conversations, trying to prepare them along the lines of: ‘Your child is very academic, but is not going to get the top grade because there’s very few of them being handed out,’ when previously these would have been straight-A* students. It will be a very emotive issue.”


www.theguardian.com/education/2017/aug/22/exam-boards-told-to-expect-surge-in-requests-for-gcse-re-marks

I wonder how the grade breakdown will compare between independent schools and publicly funded schools compared to previous years.

[Witchend]

I think article is interesting from another point of view. He says that 45% of English will be wrong up from 30%. So he's saying roughly 3 out of 10 people usually got the wrong grade (and bear in mind that could be up as well as down as they say on the ads!).

Surely rather than panicking because 45% will be getting the wrong grade this time, we should be questioning also why the 30% getting the wrong grade, I assume for some years, and he hasn't been asking for this to be addressed, and how they can do English as a fair result.

I don't think 30% of people would be near a grade boundary, so he's saying the mark system is so inaccurate it causes many marks off.

And how is this being measured? Is he looking at differences in estimated grades (in which case it would be interesting to look at subjects like maths and see the contrast there), or number of appeals (which would be difficult to compare as I think people are more inclined to appeal subjects like English than maths).

My statistics isn't brilliant, but if the scores are 9 to 1 and you gave everyone a random mark then they'd still get 11% correct.

[SoPassRemarkable]

I would imagine it's an estimate.

In old grades e.g. A* 90-100%
A70-90
B 60-70
C 50-60
D 40-50
E 30-40

All grade boundaries made up for illustration.

If everyone falls between 30-100% (I know they won't) and if we say being near a grade boundary is within 2.5% either side. You could argue that anyone getting between 30-32.5, 37.5-42.5, 47.5-52.5, etc, etc could potentially have got the wrong grade in a subject such as English.

So a lot of kids will fall close to a grade boundary. Now add to the mix more grades and more grade boundaries then there's more chance of being close to a grade boundary and more chance of an incorrect grade.

Of course this could work in your favour...

[tararabumdeay]

Old grade boundaries English Language:

A* 90 - 100%
A 80
B 70
C 60
D 50
E 40
F 30

Each grade has 30 UMS points from bottom to top. Therefore 59% of available marks would attain a grade D.

[noblegiraffe]

UMS are not the same as percentage marks.

[BoneyBackJefferson]

noblegiraffe

Oh to be a teacher of a subject that hasn't changed yet!

I am still not looking forward to what you are going through.
And the papers that we have had over the last two years have changed immensely from the starting point making past papers almost pointless.


But this is what gove wanted, a vastly reduced number of pupils that receive lower marks, he also has the get out clause that the exam boards should be more prepared even though they have had the same rubbish piled on them as the rest of us.

And the really sad thing is that it is the children that suffer.

For me the only positive is that (hopefully) most of the problems will be sorted by the time that my subject gets to where you are.

[YoureAllABunchOfBastards]

The changeover has been a nightmare, but from both sides of the fence (teacher and examiner) we have done everything we can to ensure accuracy.

Remember that any predictions about grades were guesses. No-one knew the boundaries in advance. They are set statistically after the vast majority of the papers are marked.

My worry is that whatever the boundaries are this year will be taken as gospel, and next year we will be in the same situation.

[BoneyBackJefferson]

My worry is that whatever the boundaries are this year will be taken as gospel, and next year we will be in the same situation.

Given that the drive is for marking on the bell curve, the grade boundaries will always shift to suit the cohort.

It is a fantastic way of fiddling the figures.

[noblegiraffe]

Teachers will need to remember that because of the sawtooth effect and comparable outcomes, even if next year's paper is exactly the same difficulty as this year's paper, the grade boundaries will be higher.

[woman12345]

It is a fantastic way of fiddling the figures

They could have saved time and stress by announcing the results in 2015.

After all, they changed the marking criteria of GCSE English Language half way through the final year of 2014, causing thousands to fail it then.

The Speaking and Listening component, already completed by GCSE students was removed half way through the course, so students lost 20% of their grade.

(A bit like the 25% of the grade lost to this year's OCR English Literature students who lost/are severely disadvantaged in their Romeo and Juliet grade due to an erroneous question)

www.bbc.co.uk/news/education-23884199


Tough new GCSE exams designed to bring British pupils up to a “world-class” standard will cause success rates to “plunge”, experts have warned, leaving some 360,000 pupils below the benchmark grade.

It is a fantastic way of fiddling the figures

They could have saved time and stress by announcing the results in 2015.

The EPI estimates just 40 per cent of pupils will achieve the "strong pass" grade required by the newly raised standards - a fall of 20 per cent on those meeting the benchmark standard last year.

www.independent.co.uk/news/education/education-news/new-gcse-grades-success-rates-drop-education-england-need-improvement-epi-curriculum-results-exams-a7906766.html

[BoneyBackJefferson]

woman12345

I'm not sure if you are agreeing with me or not.

What I mean by "a fantastic way of fiddling the figures" Is that it not only removes grade inflation by constantly moving the grade boundaries for a percentage fit. It means that the government can claim that they have improved the lower end as well as the pass grades there will also shift.

It also keeps us pesky teachers in place as we no longer know what the boundaries are going to be.

[woman12345]

I'm not sure if you are agreeing with me or not I do, and with above post^.

[dennishsherwood]

Here is the evidence for just how unreliable GCSE and A level grades are. The picture is a slide from a presentation entitled "Quality of marking: confidence and consistency" given by Michelle Meadows, a Director of Ofqual, at a meeting held in June 2017 (see www.gov.uk/government/news/presentations-from-ofquals-summer-series-symposium-2017).

The picture answers the question "What is the likelihood that a script in a given subject will be given a different grade if the original grade is appealed?". That's an important question, for it shows how reliable - or unreliable - the originally-awarded grade is.

As the picture shows, over each of the last four years, for GCSE physics, there has been a 15% probability that a script would be re-graded on appeal, implying that 15 candidates in every 100 were awarded the wrong grade when they opened their envelope. For English Language, 30 candidates in each 100 were awarded the wrong grade; for history, 40.

These numbers – 15 for physics, 30 for Eng Lang, 40 for history – are averages over the entire marking range from 0 to 100. Another Ofqual report shows that the probability of being awarded the wrong grade increases to 50% or more for scripts marked close to a grade boundary (see Figures 12 and 13 of "Marking Consistency Metrics", November 2016, www.gov.uk/government/publications/marking-consistency-metrics). So, if a candidate’s mark is near a grade boundary, tossing a coin might be fairer!

For English Language, the figure of 30 refers to grading according to A*, A, B.... This year, the figure could be as high as 45 candidates per 100 getting the wrong grade, as caused by the change in grading to 9, 8, 7... where 6 grades are now squeezed into the same ‘space’ previously occupied by 4, so narrowing the grade widths (and 45 = 30 x 6 / 4) – as explained on page 21 of "Marking Consistency Metrics".

These problems are a result of the fact that different markers can legitimately give the same question slightly different marks. So a script might be marked 64, or perhaps 66. If the grade boundary is 65, this is important: if the mark is 64, it’s grade C; if the mark is 66, grade B. And if an appeal is made for the script marked 64, the re-mark might be 66, and the C is up-graded to a B. The reality is that the mark given to the script is not the exact number 64, or 66: the script should be marked, say, 64 ± 2, where the ± 2 represents the legitimate variability in marking. In general, if this variability is represented as f marks (f = 2 in the example) – this being a number that can be reliably measured – then any script given a mark m ( = 64 in the example) should be represented as m ± f (64 ± 2 in the example).

The candidate’s grade can then determined not by m (= 64, grade C), but by giving the candidate the “benefit of the doubt”, and determining the grade based on m + f (= 66, grade B). If an appeal is made, then it is very likely that the re-mark does not exceed 66, and so the originally awarded grade will be confirmed. The grade will be robust under appeal, and so be reliable.

Giving the “benefit of the doubt” ensures that no candidate is awarded a grade lower than he or she deserves, no candidate is disadvantaged, no candidate loses important life chances. Nor does this idea drive grade inflation – as described in the blogs on //www.silverbulletmachine.com.

[DoctorDonnaNoble]

This is why I am frustrated by my HoD's attitude. He is just accepting some quite obviously rubbish marks (rank ordering all over the place rather than just a harsh marker).

[dennishsherwood]

A quick clarification on this chart. The vertical axis actually shows the likelihood that an appeal results in a confirmation of the original grade: so, for physics, the horizontal light blue line is at 85%, implying that 85 out of 100 appeals result in the grade being confirmed. My previous note emphasised the alternative (and much more important) view - that 15 out of 100 appeals would result in a re-grade, implying that those candidates had originally been awarded the wrong grade.

[Piggywaspushed]

The worst thing about remarking is the 'your grade can go down as well as up' blackmail. Surely all of it should be in the interests of the child, and they should be at the heart of the matter? I have always thought it should only change if it goes up (like in resits you could always reject a result). this is what terrifies parents and students out of remarks.

To give an example, my DS is 8 marks off a 6 in Eng Lang (predicted grade 7 but I always thought that was unlikely) and therefore 4 marks above the grade 5 threshold. There was a huge difference between P1 and P2 and , actually, those papers do require very similar skills. He scraped a 4 in P1. If I was his teacher, I'd want a remark for that. But I'm not; I'm his mum and am wrestling with that risk. If his lit went up to a 7 and his lang a 6 on remarks, his results would suddenly sound much less mediocre and would genuinely make a difference in uni options (in his lit , he scarped a 5 in P1 and got a high 7 in P2). I have always thought he is doomed by his own handwriting as far as some markers are concerned.

Not sure what the out and out statisticians would advise there. I have known marks to go up by 15 - 20 UMS marks one year.

[SoPassRemarkable]

Dd got 50 for paper 1 for language, 52 for paper 2. According to the grade boundaries that's an 8 in both papers. But she got a 7 overall. Some odd thing of having to multiply paper 1 score by 0.5, then add it to paper 2 and then her total score is 1 mark off an 8. This is wjec.