Sunday, 16 August 2026

Why percentages getting each grade differ between A level subjects, including French, German and Spanish

 

Thanks to David Blow, ALL Fellow,  for sharing this pdf document explaining why percentages getting each grade differ between A level subjects, including French, German and Spanish.  The explanation draws on Transition Matrices  relating to the A Levels taken in 2025. 

This information is published alongside the official DfE figures, so the 2026 information will be available late Autumn / early Spring.

https://drive.google.com/file/d/1xFt_uKwnaGyRPsuhqrPF29bIGMWW72qe/view?usp=sharing


Page 1 of the 4 page document is pasted below:

Why grade boundaries can be misleading, especially when there is a change in specification.

 

Around this time last year, I wrote this post: https://helenmyers.blogspot.com/2025/08/why-grade-boundaries-can-be-misleading.html

Below is a version revised for the context of the new GCSE specifications for French, German and Spanish in 2026.  

  • I have added a section at the start to outline the new context. 
  • The 'comparable outcomes' section has remained the same.  The GCSE specifications for French, German and Spanish  have changed, but no other subjects have changed.  The overall approach to grading is the same whether specifications have changed or not, with additional checks to ensure that candidates are neither advantage nor disadvantaged by the change.

What do we know already about the new GCSE 2026 for French German and Spanish?

In general candidates seem to have had a better exam experience this year It is hard to quantify, but the general vibe seems to have been that candidates were noticeably more positive compared with previous years when there was been widespread sense of candidates feeling demoralised as they left the exam hall.

The improvement makes sense, because in the new specification

a) questions are in English

b) exam boards are no longer required to include vocabulary not on the published specification word list.  A restricted defined vocabulary list was in place, so there should have been ‘no surprises.

This better exam experience aligns with the experience of teachers trialling sample assessments where candidates tended to achieve higher raw marks.

 However. it is really important to understand that …

 1.      Higher raw marks do NOT translate into higher grades, because that would unfairly advantage this cohort compared with previous years.

 2.      Grade boundaries are likely to be higher. As always, the conversion of raw marks to grades will vary between languages and between subjects and from year to year.

A potential concern…

The provisional entry data published by Ofqual suggests that significantly more students have been entered for Higher tier this year than previously. We strongly suspect that some schools have maintained similar patterns of entry, while others have made substantial shifts.

Until results are released, we cannot know what impact this change will have in practice. 

Comparable outcomes

In advance of the release of figures for GCSE on Thursday 20th August 2026 and to avoid wasting time and drawing misleading conclusions, it’s important to understand the relationship between raw marks and grades and how grade boundaries fit into this.  The principles apply to both GCSE and A Level.

 Ofqual have released helpful posts to explain the situation to candidates.  I have copied and pasted these at the foot of this post.

 Ofqual on grade boundaries (in 2024): https://www.gov.uk/government/news/guide-to-as-and-a-level-results-for-england-summer-2024

Ofqual 2025: https://www.gov.uk/government/publications/gcse-and-a-level-grading-what-you-need-to-know/gcse-and-a-level-grading-what-you-need-to-know

Ofqual 2026: https://www.gov.uk/government/publications/ofqual-guide-for-schools-and-colleges-2026/ofqual-guide-for-schools-and-colleges-2026#grading

 Therefore

1) Don’t look at grade boundaries in other subjects or in other languages

2) As far as any individual subject is concerned, because the national % getting each grade remains essentially constant, any change in grade boundaries will be a reflection of whether the pupils score more or fewer marks in the papers than in other years. 

It is tempting to want to talk about something which seems tangible (figures) but it can be  unhelpful when in fact the raw marks and boundaries do not tell you anything about standards from year to year,  from subject to subject , or paper to paper.

  

Copy and paste of Ofqual statements:

 Ofqual on grade boundaries (in 2024)

https://www.gov.uk/government/news/guide-to-as-and-a-level-results-for-england-summer-2024

Ofqual has said: “Exam boards have set grade boundaries this summer based on a combination of data and qualitative evidence. As in any year, there has been an important role for examiner judgement. Expert examiners have reviewed the quality of students’ work compared with previous years.

 Grade boundaries typically change each exam series and are often different between exam boards. This reflects any differences in the difficulty of the assessments. If an exam paper is harder the grade boundaries will be lower, and if an exam paper is easier the grade boundaries will be higher. This is important to ensure that students are not advantaged or disadvantaged based on the exam paper that they sit. This means that some grade boundaries are lower than 2023, while others are higher.

 Grading has taken place as normal this summer and standards have been maintained from summer 2023. Students can therefore be confident in the grades they receive, irrespective of the grade boundaries or how they compare to last year.”

 [my summary] In practice, the process of maintaining similar percentages getting each grade in each subject (subject to entry profile changes) has been in operation for nearly 15 years and is referred to as “comparable outcomes”

 Ofqual 2025

https://www.gov.uk/government/publications/gcse-and-a-level-grading-what-you-need-to-know/gcse-and-a-level-grading-what-you-need-to-know

 Grade boundaries change from year to year 

If an exam is easier than in previous years, the grade boundaries for that paper will be higher. If it is harder, the grade boundaries will be lower. 

The difficulty of exam questions varies year to year, even though exam boards try to keep the level of demand consistent. That’s because it is impossible to determine how difficult students will find a paper until it is taken. 

This is why new grade boundaries are set each year – to reflect the difficulty of that particular paper, and to ensure that it is no easier or harder to get a grade in any given year.  

 The standard of work needed to get each grade remains comparable year on year 

The quality of a student’s work in their exams determines the grade they get and no-one is advantaged or disadvantaged by the year in which they happen to sit the exam​. There are no quotas for the number of students who must pass or fail a subject each year, and no cap on the number of students that can get each grade.  

The fact the number of top grades awarded is relatively stable from year to year reflects the fact that student performance is typically relatively stable too. ​ 

Grade boundaries are set with the aim of maintaining standards over time, not to ensure that the same proportion of students get a particular grade. 

 Grade boundaries are decided after students take exams and when marking is nearly complete 

Exam boards set grade boundaries by analysing data about the students taking the qualification and comparing it with those that have taken it previously. This data is used to identify potential grade boundaries, which reflect whether students found the exam slightly harder or easier than the previous year.  

Senior examiners then scrutinise samples of work from students who have achieved marks on and around these potential grade boundaries. This is compared with work of students from the previous year who were on the same grade boundaries.  

Those scripts are studied to determine if the standard of work at the potential grade boundary is appropriate. Senior examiners use the evidence from their scrutiny in combination with the data, to recommend final grade boundaries. 

 The number of students achieving each grade can differ between exam boards offering the same qualification 

Grade boundaries typically differ between exam boards, because grading reflects the level of difficulty of each particular paper. All exam boards must cover the same subject content and assessment objectives, and follow the same rules. But there are often small differences between the type of questions they ask or the balance of different types of questions in a particular paper. 

Similarly, while exam boards try to keep levels of demand consistent year-on-year, this is also very difficult to do in practice.  

Where grade boundaries increase between one year and the next, or are higher with one board than another, it does not mean it is more difficult to achieve a grade. Rather, it reflects that that exam paper was easier. Equally, where grade boundaries are lower with one exam board than another that reflects that the paper was harder, not that it is easier to get a particular grade. 

 Ofqual 2026: https://www.gov.uk/government/publications/ofqual-guide-for-schools-and-colleges-2026/ofqual-guide-for-schools-and-colleges-2026#grading

Grading for GCSEs, AS or A levels in 2026

Ofqual oversees exam boards to ensure that it is no more difficult to achieve a grade in a subject from one year to the next, and between different exam boards. This means that the standard of work to achieve a particular grade in 2026 will be comparable to summer 2025.  

Grade boundaries often vary year to year, to reflect fairly any changes in how difficult the paper is. The quality of a student’s work in their exams determines the grade they get and no-one is advantaged or disadvantaged by the year in which they happen to sit the exam. 

There are no quotas for the number of students who must pass or fail a subject each year, and no cap on the number of students that can get each grade. The proportion of students receiving different grades does vary a little each year. But the standard of work required to receive a particular grade remains comparable.   

More information about how grading works in GCSE, AS and A levels can be found in Ofqual’s Understanding grading toolkit for schools and colleges.

And finally .... please join ALL if you are not already a member!

ALL is extremely grateful to the volunteers who carry out work freely on behalf of all language teachers. There is, of course, a cost to the Association in running  services - coordinating activities, managing website content, organising events and webinars - and this is funded by member subscriptions.

Many colleagues benefit from ALL’s analysis, advocacy and resources without realising that these are produced by volunteers who themselves pay to be members. In an era when people increasingly expect everything “for free”, it is worth remembering the value of belonging to a professional organisation that works year‑round to support language education nationally.

I encourage those who are not yet members, but who benefit from our work, to join ALL - whether as individual members (giving you voting rights and personal copies of our publications) or as group members. Join here: https://www.all-languages.org.uk/join/become-a-member/

 

 


A Level MFL results 2026

JCQ

I attended the JCQ briefing for stakeholders on Thursday 13th August on behalf of The Association for Language Learning (ALL).  

Each year the quality of presentation, preparation and documents improves, and there were many expressions of thanks at the end.  It is extremely helpful to receive the data an hour in advance of the meeting (-there is a strict embargo until 9.30 am) as this gives us the chance to  check any hypotheses we may have formed and to prepare relevant questions.  In particular this year, I commend JCQ on two new aspects:

(1) The new page Exam results data from which you can easily access current and past data by selecting the year of interest: https://www.jcq.org.uk/exam-results-data/

(2) The provision of a grade breakdown for each of the languages within the overall category ’Other Modern Languages’ in GCE Miscellaneous, 2026 (Table 3, Tab 3d): https://www.jcq.org.uk/wp-content/uploads/sites/2/2026/08/Table-3-GCE-Miscellaneous-2026.xlsx  This has been very helpful re: AL Chinese which we are continuing to follow up with Pearson.

 Thank you, JCQ!

ALL follow up work

ALL has issued a statement which gives key data, presidential comments and links.  https://www.all-languages.org.uk/news/all-statement-on-2026-a-level

Over the coming weeks, we will follow up issues or concerns expressed by our members.  It is invaluable to have direct contact with our corporate members, especially at this time, including exam boards and representatives (including data specialists) who are keen to engage with us and to offer clarification or pursue concerns further.

Extract from statement: Analysis

I have copied and pasted below elements relating to analysis which may be particularly useful to teachers and managers analysing their own results.  Many thanks to David Blow, ALL Fellow, whose spreadsheets make it easier to quickly extract percentages and raw numbers from the tables and track them from year to year.

The number of students studying A level languages continues to be a matter of grave concern. Whilst the overall number of A level entries for all subjects has increased substantially since 2019 (906,425 v 804,851) the total number taking French, German or Spanish has actually declined (17,616 v 20,013; 1.9 % of total v 2.5%).  

  • The overall number of entries for all subjects has increased by 2.9% since 2025.  
  • The increase in German entries, which exceeds this, is welcome, but it is still down by 25% compared with the 2019 entry numbers.  
  • Although Spanish and French entries have increased by 1.3% and 0.4% respectively, these rises are lower than the overall increase. 

This raises concerns about the UK’s ability to produce enough qualified linguists for the country’s needs. It is vital that steps are taken to ensure students get a broader and richer post-16 experience including a language.

These low numbers studying French, German and Spanish at A level reinforce the need for the continuing severe grading in ML at GCSE to be addressed by Ofqual and DfE so that more students are encouraged to continue with a language to GCSE and then on to A level.  Systematically, arising from an historical anomaly, students on average get half a grade lower in their GCSE ML compared with their other EBacc subjects, leading to a false perception that they are not as good at ML as at other subjects.  They are therefore less likely to take it at A level. We appeal to all stakeholders to keep raising this matter, especially in this year when there are new GCSEs specifications being examined for the first time in French, German and Spanish. It is important to remember that however pedagogy and the nature of exams may change, the approach of comparable outcomes means that severe grading will persist.

Appendix with detailed numbers

The overall number of A level entries in all subjects has increased by 2.9% to 906,425 in 2026 from 882,509 in 2025, having decreased by 0.5% from 886,514 in 2024. The 2026 overall entry figures have increased from 804,851 in 2019 (13% increase).

French entries have increased by 0.4% to 6,888 in 2026 from 6,858 in 2025, having decreased by 9.8% from 7,544 in 2024.

German entries have increased by 3.1% to 2,294 in 2026 from 2,224 in 2025, having decreased by 8.5% to 2,224 in 2025 from 2,431 in 2024.

Spanish entries have increased by 1.3% to 8,434 in 2026 from 8,325 in 2025, having increased by 1.1% to 8,325 in 2025 from 8,238 in 2024.

‘Other Modern Languages’ entries have increased by 4.1% to 6,944 in 2026 from 6,672 in 2025, having increased by 3.8% to 6,672 in 2025 from 6,429 in 2024.

This is all in the context of a demanding subject such as Further Mathematics having over 21,000 entries.

Comparing entry numbers between 2026 and 2019 (pre-pandemic) there is a clear decline. French has dropped by 17.6%, German by 24.4%, Spanish by 2.2% and ‘Other Modern Languages’ by 16.2%.  By contrast, in the same period, Economics has risen by 53% to 47,157, and Psychology by 22.6 % to 79,201.

These numbers are taken from the JCQ UK cumulative figures published Thursday 13th August 2026.

Appendix with links

JCQ has created a new page named ‘Exam results data’ from which you can access current and past statements and data by selecting the year in which you are interested. https://www.jcq.org.uk/exam-results-data/

Direct links to the press notice (pdf) and 3 results tables relevant to Modern Languages (Excel spreadsheets) are as follows:

(3) Level 3 Press notice 2026: https://www.jcq.org.uk/wp-content/uploads/sites/2/2026/08/L3-Press-notice-2026.pdf

(4) A level by subject age sex country, 2026 (Table 1): https://www.jcq.org.uk/wp-content/uploads/sites/2/2026/08/Table-1-A-level-by-subject_age_sex_country-2026.xlsx

(5) AS level by subject sex country, 2026 (Table 2): https://www.jcq.org.uk/wp-content/uploads/sites/2/2026/08/Table-2-AS-level-by-subject_sex_country-2026.xlsx

(6) GCE Miscellaneous, 2026 (Table 3) [Note that this includes data about ‘Other Languages’ in Tab 3d]: https://www.jcq.org.uk/wp-content/uploads/sites/2/2026/08/Table-3-GCE-Miscellaneous-2026.xlsx

(7) A level interactive results dashboard: https://www.jcq.org.uk/a-level-interactive-results-dashboard/

Ofqual, Pearson – A level Chinese.

https://www.gov.uk/government/news/ofqual-fines-pearson-2-million-for-rule-breaches-affecting-thousands-of-students

Ofqual has shared a range of comments and analysis as follows (links embedded):  


Please join ALL! https://www.all-languages.org.uk/join/become-a-member/

ALL is extremely grateful to the volunteers who carry out this work freely on behalf of all language teachers. There is, of course, a cost to the Association in running these services - coordinating activities, managing website content, organising events and webinars - and this is funded by member subscriptions.

Many colleagues benefit from ALL’s analysis, advocacy and resources without realising that these are produced by volunteers who themselves pay to be members. In an era when people increasingly expect everything “for free”, it is worth remembering the value of belonging to a professional organisation that works year round to support language education nationally.

I encourage those who are not yet members, but who benefit from the work of ALL, to join us - whether as individual members (giving you voting rights and personal copies of our publications) or as group members. Join here: https://www.all-languages.org.uk/join/become-a-member/