I read an article from Cambridge published by the Higher Education Policy Institute.  It’s a curious read.  The salient part is this:

“The proposed introduction of a Home Office English Language Test (HOELT) raises the stakes still further. The Home Office has indicated an interest in at-home invigilation. While innovation of this kind may appear to offer greater convenience, it also risks undermining quality, fairness and security. The HOELT process must be grounded in evidence, setting high minimum standards and ensuring robust protections against misuse. High-stakes decisions such as the creation of HOELT should not be driven by cost or convenience alone. They should be driven, instead, by whether the system enables talented students to succeed in the UK’s competitive academic environment, while safeguarding the country’s immigration processes.”

This is a fair and principled stance.  And sincerely held, I’m sure.  But to go the route of telling the Home Office that the test they are desirous of (a fully remote at-home test) is the wrong test is a curious move.  It could backfire.  I’m imagining a restaurant-goer who orders a salad only to be told by the chef that he ought to get a steak instead.

What’s more curious is that Cambridge has been offering an at-home IELTS test for more than five years. And with nary a peep about security concerns.  Indeed, to my eye it seems to have more reliable security than the traditional test center administration.  Hundreds of people take this version of the IELTS every day. So why not lean into that?

Duolingo’s bid is completely beatable.  Even by the standards of at-home testing Duolingo’s approach is new and scary.  I can picture the HOELT tender being won by a testing firm that openly states its love for at-home testing but presents the sort of conservative game plan that the Home Office is probably looking for.

A result where a smaller testing firm emerges to snag the tender might be Cambridge’s worst nightmare.  And a source of great amusement for the Duolingo team.

The Daily Sun (Bangladesh) has published a long article on IELTS exam leaks and organized cheating. I’ll post a link in the comments.

According to their report, two suspects have been arrested in Dhaka for their part in a cheating ring.  The Sun reports that the suspects used leaked IELTS tests to prepare students for the test.

Says the article:

“Students who followed the racket’s instructions reportedly achieved their expected band scores, raising serious concerns over the integrity of the IELTS examinations in the country.”

This comes after a lengthy Daily Sun investigation of a separate IELTS cheating ring operating on the same principle.  Basically, it is alleged that ringleaders acquired test forms the night before specific administrations of the IELTS test from a leaker operating within the IELTS partnership.  After that, customers were (allegedly) secreted away to local hotels where they had 4.5 hours to memorize the correct answers.  The next morning they (allegedly) took the test at a variety of test centers across the city.

Says the article:

“On 25 April, a correspondent stayed at Hotel Afford Inn in Uttara, where around 100 students were lodged overnight. Leaked questions were reportedly handed out at mid-night, after students were separated into teams according to their exam centres. The next morning, the students were transported by minibuses, microbuses, and CNGs to [test] centres including Patronas at Panthapath, Compass at Banani, IALC at Dhanmondi, and Penstone at Uttara. A similar operation was observed on 23 May at Hotel Central Inn in Motijheel, where 120-130 students were being prepared for the following day’s exam.”

Beneficiaries of this scheme confirmed (allegedly) that the leaked questions were identical to those on the real IELTS.

The Daily Sun report suggests that the aforementioned cheating rings “highlight serious flaws in exam security.”

The security of the paper-delivered IELTS has come to the forefront in recent months. Note that the IELTS Partnership pulled the plug on paper delivery of IELTS for UVKI in Bangladesh at the end of July.

The paper-delivered IELTS has also been eliminated or restricted in key markets of Vietnam, Pakistan, Sri Lanka, Malaysia and Nigeria. Meanwhile, test takers globally have recently been forbidden from accessing the paper-delivered test outside of their country of residence (or country of citizenship).

Some have speculated that these moves are connected to longstanding security concerns about the paper-delivered IELTS.

Earlier this year the UK Home Office raised concerns about possible cheating on IELTS tests, citing comments from HEIs that IELTS scores “don’t necessarily replicate the reality of the applicant’s English language ability.”

If you are still reading, do take a moment to peruse the whole Sun article.  It contains the most succinct description I’ve ever read of the ol’ paper-delivered IELTS cheating technique (allegedly).

Here’s a video from the Australian Broadcasting Corporation (ABC) about English tests in Australia and recent changes. Among other things, it touches on the fact that many visa holders in Australia test and re-test and re-test and re-test. Indeed, the recent financial results from IDP Education Ltd touched on how this kind of on-shore testing is a growth area for the IELTS.

My most recent “picture of the week” over on Substack was a shelf of test prep books at the Alderney Gate branch of the Halifax Public Library.  A fun way to do some boots-on-the-ground research into testing trends is to see what test prep books are checked out of your local library.

For instance, right now all six copies of IELTS 18 (General Training) are checked out of Halifax’s library system, with a waiting list of 15 holds.  Four out of five copies of IELTS 18 (Academic) are checked out.

Meanwhile, over on the left coast all 22 copies of IELTS 18 (General Training) are checked out of the Vancouver Public Library system, with a waiting list 7 holds deep.  Twenty-two out of twenty-four copies of the Academic version are checked out.

As IDP Education has indicated a few times in recent weeks, on-shore testing in Canada is a pretty big deal nowadays.  That’s partially because of new testing requirements for PGWP applicants that came into force back in November.

Interestingly, there seems to be less demand for IELTS 19.  The new cover design might have some people thinking it is a third-party product.

IDP Education Ltd has released Full Year Financial results for FY 2025.

IELTS volumes are down 18% for the year for a total volume of 1,293,800 tests.  Last year’s volume was 1.5 million tests.  The peak for IDP was 2022, when 1.9 million tests were administered.

IELTS volumes are down 50% in India for the year.  This comes after a 42% drop in India back in 2024.

Outside of India, IELTS volumes are up 2%. This is attributed to onshore testing in Canada and Australia.

IDP failed in its goal to launch IDP IELTS in China in the fiscal year.

Shares are up 31% on the news.

IELTS test statistics for 2024/25 are now available (I think they were published earlier this month). Interested readers can now access updated demographic and test taker performance data. Nothing unusual caught my eye, but let me know if you see something interesting.

Writing in the Financial Times, Cambridge University Press & Assessment’s chief product officer Pamela Baxter notes that “international students, who are expected to demonstrate only minimum standards of English for their admission to UK universities, often struggle to keep up with coursework and to contribute usefully to classroom discussions.”

This, she points out, “is a disservice to those students and their peers.”

Baxter ascribes this problem, in part, to the acceptance of English tests that rely too heavily on AI.  As proof, she cites a study (published in the ELT Journal) that the IELTS partners have promoted quite a lot since it was published at the beginning of last month. We’ve heard this message ad nauseum from the IELTS partners in recent weeks.

But one is left wondering about the connection that has been made between AI-based tests and poor student performance.  When the study was carried out, the Duolingo English Test was accepted at just 6 out of 40 surveyed higher-ed institutions.  Another test singled out by the authors of the report, the Oxford ELLT, was accepted by so few institutions that it wasn’t even listed among the results of the survey.

These tests have increased in popularity a bit since the study was completed, but not by much – the Duolingo Test was accepted at 5 out of 20 institutions I surveyed earlier this year.  The Oxford ELLT Test was accepted at 4 out of 20.

So why are we blaming tests that very few schools actually accept?

Traditional human-centric tests like the IELTS, TOEFL and Cambridge C1 tests meanwhile, were (and are) accepted at almost all major HE institutions in the UK.  Accordingly, such tests are still used by the vast majority of international students heading to the UK who are required to submit scores in the first place.  Odds are high that many or most of the struggling students referred to by Baxter used traditional human-centric tests during the admissions process.

This is not to say that traditional tests are doing a poor job of measuring the language skills of students heading to the UK.  Nor would I suggest that such tests are failing to contribute to positive washback among the test taking problem.  The tests are doing their job.  They are measuring English fluency. My point here is that if students are struggling to use English in the classroom, there are problems that go far (far, far, far) beyond one’s choice of English test.