Exam Data Nerds step this way

[noblegiraffe]

A bunch of tools for looking at the GCSE data here:

analytics.ofqual.gov.uk

My favourite is this one: analytics.ofqual.gov.uk/apps/2018/GCSE/9to1/
You can look at the distribution of grades for all the reformed subjects which is quite cool, but you can also look at the distribution of grades in a certain subject for pupils who got a particular grade in a different subject or two.

Want to know how many students got 9s in Lit, Lang and PE? It’s 220 out of the 75,740 students who took that combination, so only about 0.3%.

And look at the grade distribution for Ancient Languages!

Also surprised to see that the modal grade for English Language was a 3.

So much fun to be had here.

On a more dispiriting note, 2000 students got a U in both English and Maths. After 11 years of schooling, to be entered and to fail both

The ancient languages graph is vastly different from the other subjects! That is so telling! Big standard schools don't tend to offer this subject!

*bog standard

Interesting. Thanks for sharing.

It is interesting. I wish I knew and understood more about data and statisics. I have actually been in tears about the results of my dept tonight, bit knowing the modal score for Eng Lang was 3 makes me feel a little better. That is one of the things I hate most about the system - hearing other students have done badly/that the same students have done worse in other subjects makes one feel a little better - until the guilt kicks in of course.

If the mode is 3, then why is that not reflected in the headline stats that will soon come out?

That's really interesting. Thanks for sharing.

Ah just I’ve just done a bit of digging. The modal score from the data in the Ofqual link is 3 for English, but that includes resits, which are notoriously terrible.

When you look at the data for 16 year olds only, the mode is 5, but only just. 19.6% of 16 year olds got a 5, but 19.2% got a 3. That’s still a lot.

Breakdown of results by country within UK, by age and gender are here analytics.ofqual.gov.uk/apps/2018/GCSE/9to1/ - it’s a big document so if you can’t find what you want immediately, just keep scrolling.

Dickybow, a couple of years ago when ofqual were threatening to cap numbers of 8s and 9s to the same percentage for all subjects ancient languages were among the most vociferous protestors, arguing that they had a uniquely able cohort. They won on that basis, as you can see by the graphs.

Capping the ancient language grades would kill them overnight I think. I’m on my phone, but if the data is all that Noble says it is, it should be easy to test the assertion by comparing the average “grade achieved in all other subjects taken” grade for students taking Latin/Greek with the equivalent measure for, say, Spanish, Geography or Art.

Using the comparison tools the ancient languages kids did amazingly in most subjects

Yes that’s what I expected. Most of the Latin and especially Greek candidates are expecting to get straight 7/8/9s. If you marked the ancient languages to produce the same grade distribution as the whole GCSE population then you’d be seeing a lot of those children getting a random 4 or 5 in amongst their 7+s and any school which gives a damn about league tables would suddenly find them a highly unattractive option.

Just under 6,000 students got a U in English Lang. Most of those got a 1 or U in maths, none got higher than 2. I feel for my students with SEND.

Over 11,000 students got a U in maths. Why does the maths have almost double the fail rate of English?

Poor teaching over many years? English is used in essay subjects. (Or when writing a few paragraphs). So gets more practice. Maths is less widely used in other subjects.

Surely some children are not going to be able to do well at these exams? It was ever thus. Not every child has the intelligence to get above a 2 or the staying power to learn. It would be most revealing if we knew where these children go to school. Are they in special schools? There is a good chance they are not in mainstream education or are in a pru or in a unit in a mainstream school. Therefore they may well have learning difficulties. Therefore it’s unfair to compare them with others. They may have done well to get a 1 or 2 grade.

It would probably be unfair to have a bell curve or wotnot for Ancient Languages. It could mean that someone on 90% could get a U if everyone else got 91-99%.

Dd got 94% in Maths in her Year 10 mocks and was told that that might be a Level 7 depending on where the 8 and 9 grades were set. I thought that odd

I agree bubbles, it wouldn't be fair to compare

Also agree with the comments about a bell curve graph for ancient languages being unfair. In fact it makes me think any system where the top 3%get a certain grade regardless isn't great. If one year 50% was the highest mark its not fair that the year after that it could be 98% (extreme example but theoretically could happen!) An employer wouldn't be able to tell that candidate A is not as competent in the subject as candidate B as they got the same grade.

Irma that's bc teachers were unnecessarily scared about this grade 9 thing. A 9 turned out to be 84% this year.

Well that was fun!

I'm a data nerd, and work with my school's data, so good combo?

Ugh. Hit post too soon.

I didn't really get to dig into the GCSE numbers yet this year as I instantly went on holiday come Thursday, but a lot of our students who had Us in both maths and English either were EAL and relatively recent arrivals, had SEN, or were the sort who turned up but slept through the whole exam.

The whole thing where there are no benchmark skills and 40% of students are doomed to failure (per subject with exceptions as above) because of grade boundary malarkey is truly depressing and I weep for this country and its stupid examination system.

Fascinating. This will be very useful.

Does this data include iGCSEs?
When ascertaining grade boundaries, say for Marhs, are they done for all marhs iGCSE & GCSEs or by individual boards etc. I’ve heard there’s disparity between the boards on difficulty etc so wonder if grade boundaries are determined within each board or all thrown into the pot? Not too sure what the fairest way is either.

If you want exam statistics by exam board or data for IGCSE, all the possible links you could want are in a handy document linked on twitter:

twitter.com/missdcox/status/1032997604676825088?s=21

Exam boards set their grade boundaries individually, but these are checked by Ofqual, they had an argument with Edexcel a couple of years ago about where to set the C boundary for maths and made them change it. Don’t know if they regulate IGCSE.

We did have a number of Us this year : more for Lit than Lang. Not actually SEN students . They tend to do a different qualification in reality. The Us tend to be those we worry most about : pre NEET. Often disadvantaged, certainly disenchanted and disenfranchised. Dis everything in fact...

This new system does nothing for them. No coursework and no contribution from speaking and listening, so they are massively affected by persistent absence, for example, and the need to study hefty texts pretty much ad nauseam -stimulating for many and a trun off for many, too.

I like the breakdown by county pages!