One Listening Score on the UK Visa List Is Lower

The UK's approved test lists are two spreadsheets, four CEFR levels and five providers. Almost every cell follows the same pattern.

One does not, and it sits at the top of the scale.

SpeakShark is free for the skill nobody should be planning around an anomaly for, three AI conversation sessions a day with no card.

In this guide: the cell · in both files · why it stands out · two explanations · do not rely on it · the other slip · what to do · limits · method · FAQ

Key takeaways

• LanguageCert International ESOL SELT at C2: listening 25/50, everything else 33/50.
• The same test requires 33/50 in listening at B1, B2 and C1.
• The figure appears identically in both official lists.
• We are not calling it an error, and not calling it correct either.
• Do not plan around it. Aim at 33/50.

The cell that breaks the pattern

Here is the LanguageCert International ESOL SELT four skill test, exactly as the approved lists give it.

Fifteen cells reading 33 out of 50, and one reading 25.

Read literally, an applicant proving C2 on this test needs a lower listening score than an applicant proving B1 on the same test.

It appears in both official lists

This is why the finding is worth publishing rather than shrugging at.

The Home Office maintains two separate documents: a list of approved tests taken in the UK and a list of approved tests taken outside the UK. They are different files with different contents, since two of the five providers appear on only one of them.

The 25 out of 50 figure appears in both, in the C2 row, with the other three skills at 33 out of 50 in both.

A number that appears once could be a slip in one file. A number that appears identically in two separately maintained documents has been carried, or copied, or is intended.

Why it stands out so sharply

Every other numeric test on these lists behaves consistently across its four components at every level.

Seventy-odd numeric cells across the two lists, and one of them is out of step with its own row and with its own column.

That is what makes it visible at all. In a table where thresholds varied by skill, 25 next to 33 would be unremarkable. In a table where they never do, it is the only thing on the page that does not fit.

The band tables it sits beside are covered in IELTS for UKVI by CEFR level, PTE Academic UKVI scores by CEFR level and LanguageCert Academic against General SELT.

Two explanations, and no way to choose

We can see two, and the published documents do not let us pick.

It could be intended. A raw score out of 50 is not a percentage of ability. If the C2 listening paper is substantially harder than the C1 one, a lower raw score can represent a higher standard. Scales do behave that way, and the same logic explains why TOEFL's raw sections map to bands on different curves, covered in TOEFL's new 1 to 6 scale.

It could be an error. Long-standing transcription mistakes survive in maintained lists precisely because nobody rereads a row that has been there since 2020, and copying between two files propagates rather than corrects.

Nothing on the row, and nothing in either document, distinguishes the two.

We are reporting it as published. That is the whole of our claim, and it is deliberate: an earlier round of this research recorded the figure with the instruction to report it as it appears in both lists and not to build an explanation on top of it.

Why you should not rely on it

Whatever the cause, the practical advice is the same and it is not the fun one.

An anomalous published figure is the worst possible thing to plan an application around, because the two outcomes are asymmetric.

Aim at 33 out of 50, which is what the same test asks at every other level, and the question stops mattering.

This is the same reasoning we applied to a live discrepancy in the American federal requirements, where two published pages give different PTE speaking figures and the safe move is the higher one, covered in every US healthcare English test now ranks speaking highest.

The other slip on the same list

While reading these files closely, one plain typing error is visible and worth a sentence.

In the list of approved tests taken in the UK, the A2 row for Trinity College London reads Graded Examinations in Spoken English (GESE) Grace 3, where every other Trinity row reads Grade.

Nobody is going to be misled by that. It is included here for one reason: it establishes that these documents contain uncorrected typing errors, which is relevant context when weighing whether a numeric cell is intentional.

We covered what GESE actually is, and how its grades map to CEFR levels, in Trinity GESE as a UK visa test.

What to do with this

• Aim at 33 out of 50 in every skill, including C2 listening.
• Do not build a plan on an anomalous cell.
• Check the current list on the day you book. Both files are revised.
• Read the row for your own level, not the row above or below it.
• Confirm with the provider if a published figure looks wrong to you.

Whichever number applies, the component candidates most often miss is speaking, not listening, and it is the one that does not improve by reading quietly. Role play scenarios and a daily speaking partner are the practical routes, testing your speaking level free shows where you stand, and CEFR levels for speaking explains what the levels describe. What C2 actually means covers the level in question, the UK routes that only test speaking and listening covers the two skill alternative, and English certificates that expire covers validity. For how examiners mark spoken English, Cambridge speaking criteria by level works from real scored candidates.

What we could not verify

We did not establish whether the figure is intended. Both explanations remain open.

We did not ask LanguageCert or the Home Office. This is a reading of published documents.

We did not verify the scale behind a score out of 50. What 25 or 33 represents in ability terms is not published on these lists.

We did not check historical versions of the lists. The figure carries an effective date of 3 April 2020 in both files and we have not traced it back further.

The lists are revised. Anything here may be corrected after we published.

This is not immigration advice. It is a reading of two published government lists, both linked below.

How we researched this guide

Anomalies are only visible against a pattern, and the pattern here took reading both spreadsheets as data rather than looking up one row.

Somebody checking the test they intend to sit reads a single line and never sees that every other numeric cell on the list follows the same rule. Parsing all of it and comparing rows to rows and columns to columns is what makes one cell stand out.

Checking the second file was the step that turned it from a curiosity into something publishable. If 25 appeared once, the honest note would be that one file contains a slip. Appearing identically in two separately maintained documents is a different fact, and it is why this guide exists.

We then stopped, on purpose. An earlier round of this research wrote down the instruction to report the figure as published in both lists and not to build a claim on why, and that instruction was right. There is a version of this guide that confidently explains the number, and it would be fiction.

The Trinity typo went in for one reason: it is evidence that these documents carry uncorrected errors, which is exactly the context a reader needs to weigh the numeric cell for themselves. It is not evidence that the numeric cell is wrong.

Practise speaking, from SpeakShark

SpeakShark is an AI English speaking practice app, and the practical advice in this guide is the same as everywhere else in the series: be comfortably above the line rather than on it.

An anomalous cell, a scale change, a regulator restating a threshold, a list being revised. All of those are risks to somebody sitting exactly on a requirement and irrelevant to somebody a step above it.

You talk, the AI answers what you actually said, and you get speaking feedback while the conversation is still running. The free tier gives basic feedback; the detailed pronunciation and grammar breakdown is on Premium.

Being straight about the limits: a free session runs five minutes with four turns, we issue no certificates, and we are not affiliated with LanguageCert or any approved SELT provider.

Start free with three sessions a day and no card. Paid sessions run ten minutes with unlimited turns. Limits are on the pricing page, and how it works walks through a session. Sign up here if you would rather clear a threshold than argue about it.

We are a speaking improvement tool. We are not an exam preparation provider, and we are not affiliated with Cambridge English, the IELTS partners, Pearson or any exam board. For test format, booking and official practice material, go to the exam body directly. Use SpeakShark to make your spoken English stronger, and use official material to learn the test.

• Site: speakshark.com
• Support: speaksharksupport@gmail.com
• Free tier: 3 AI conversation sessions per day, no card

Sources

• GOV.UK, list of approved Secure English Language Tests taken in the UK: the LanguageCert International ESOL SELT rows at B1, B2, C1 and C2 including the C2 listening figure of 25 out of 50, and the A2 Trinity row reading GESE Grace 3
• GOV.UK, list of approved Secure English Language Tests taken outside the UK: the same LanguageCert International ESOL SELT rows with the identical C2 listening figure
• GOV.UK, prove your English language abilities with a secure English language test: the guidance the two lists sit under, including the two year window and the approved location and provider conditions
• Council of Europe, CEFR Table 1, common reference levels global scale: the definition of C2 that the row is mapped to

FAQ