The Quantum Hub and Duolingo English Test have published a white paper about the DET. The contents will be familiar to most regular readers, but it is worth checking out if you are still puzzled about how the test was able to become so popular over the past five or six years. There is a tendency to assume that the DET is popular because test takers think it is really easy. But it goes deeper than that. Ignore those deeper reasons at your own peril.
Category: Duolingo
The University of British Columbia now fully accepts the Duolingo English Test for undergraduate admissions. Previously, the school’s website still mentioned that it was acceptable only in cases where no other test could be taken.
I think Duolingo has now achieved near-universal acceptance among undergraduate programs in Canada.
The University of Victoria will accept Duolingo English Test scores starting in May of next year. I think they were the largest non-accepting school in Canada last time I did a scan of Canadian institutions.
Duolingo published Q3 results last week. Revenue from the Duolingo English Test amounted to $9,649,000 for the quarter. If we divide that number by the cost of taking the test ($70) we get a total of 137,842 tests for the quarter. That’s a decline of 17% compared to the same quarter last year. This figure tracks with what we’ve seen so far this year: Q1 was down 17% and Q2 was down 12%.
Just keep in mind that this is a rough estimate, as some people pay less for the test by purchasing a bundle and others pay more for the test by purchasing express scoring. Others pay nothing at all by using a waiver from the Duolingo Access program (25k waivers were distributed last year).
How does this compare to the industry as a whole? By my data, PTE volumes were down 10% in the first half of 2025, while IELTS volumes at IDP were down 11% for the first half of 2025.
IDP will publish a new half-year report in February.
We might learn about IELTS volumes at the British Council if they publish their annual return in January, as required (but they probably won’t). ETS doesn’t publish TOEFL test volumes, but keep an eye out for the release of the organization’s annual audit in early January.
Between June 2021 and June 2022, IDP Education administered about 1,900,000 IELTS tests. During that same period, Pearson administered about 591,000 PTE tests.
Between June 2024 and June 2025 IDP administered about 1,293,800 IELTS tests (a 600k drop). During that same period, Pearson administered about 993,000 PTE tests (a 400k increase).
This according to the annual reports published by each firm.
Apples and oranges, I suppose. But one is left to wonder when the number of tests administered by Pearson will surpass the number administered by IDP education. One is also left wondering if Pearson (which seems to focus on a narrower range of markets) already does more tests in India than IDP.
(By my crude calculations, DET jumped from 494k to 684k in the same time frame)
I finally got around to experiencing the new room scan for the Duolingo English Test. The room scan is mostly painless. It resembles that of the British Council’s EnglishScore test: the test taker just has to slowly spin around their room while holding the secondary camera (their phone). There is also a second scan, which hasn’t been discussed much in this space. To complete it, the test taker first points the phone at their keyboard, and slowly moves it towards the space behind their computer. It is a little clunkier than the main room scan, and the test taker might attempt it a couple times before doing it properly.
Some have argued that requiring test takers to have a compatible smartphone is burdensome. This is true, to a certain extent. That said, many test takers will find that it alleviates the burden of completing room scans by carrying a laptop around their apartment. Remember that most test takers don’t have a cool-guy Macbook like the people who design and market tests. Instead, they have a barbell of a Windows machine with a bunch of peripherals connected via USB and a battery that has to be plugged in all the time because it maxes out at a 5% charge. Doing a room scan with such a machine really sucks. Ask me how I know.
The IELTS partners have published a new guide to assessment literacy and picking English tests. It’s pretty good.
They suggest asking five questions:
- Is there research to support test validity?
- How are speaking and writing assessed?
- How do your tasks align with academic demands?
- What security measures are in place?
- Is it possible to review how test taker performance is assessed against specific criteria?
These are all excellent questions. Score users should certainly ask them. One might also read Goodine’s Guidelines.
From a business perspective, these questions seem to highlight how IELTS Official continues to find itself in a tight spot. They are dealing with competition from more contemporary tests like the PTE Academic Test and Duolingo English Test which are successfully putting forth the argument that shorter items (what IELTS calls “limited-response items”) have a role to play in snapshotting the language proficiency of an applicant (even if only in combination with longer items). On the other hand, they also face competition from fairly traditional tests like LANGUAGECERT which, as recent events have suggested, may be supported by much more stringent security measures than is the IELTS.
I saw in this Youtube video that the Duolingo English Test now requires a room scan. The required scan is a 360 degree motion (with a phone) akin to British Council’s EnglishScore test, plus a look at the space behind the computer.
I’ve never been a fan of room scans. I find them cumbersome and I’m unconvinced of their effectiveness. For what it’s worth, this one seems a bit less obtrusive than most of the scans I’ve carried out. It seems unlikely to result in a request to drag furniture around my house or to take down all of my Stone Cold Steve Austin posters.
I’ll take the test later this month and report back with details.
One wonders if this nod to more traditional at-home testing is a reflection of Duo’s new status as a co-frontrunner for the HOELT tender. Note also the reminder in this week’s THE article that every DET test is reviewed by a human proctor (something that not everyone is aware of).
The ETS Research Institute has published a new study which compares the TOEFL iBT and Duolingo English Test (DET). It is dated April 2025 but is new to me and it appears to be new to the Internet. The study was carried out by Sara Cushing of Georgia State University.
Curiously, this study compares the DET to the current incarnation of the TOEFL iBT rather than the heavily revised version which will debut in January. One is left wondering if Professor Cushing was even told that she was researching a test that would cease to exist a few months following publication. Surely her time could have been better spent comparing the DET to the revised TOEFL.
The study is very critical of the DET. The author levels pretty strong criticisms at the test and the choices made by its designers.
Interestingly, many elements of the DET which are singled out for criticism will soon be found in the new TOEFL. Meanwhile, many aspects of the TOEFL which are specifically praised in the study will not be found in the new TOEFL.
For instance, Cushing praises the TOEFL for the inclusion of integrated speaking and writing tasks that reflect the environment students will find themselves in at university, for its use of lengthy reading and listening passages, for its exclusive focus on academic content, and for eschewing short item types that test low-level fluency.
In contrast, the DET is criticized for its lack of integrated speaking and writing tasks, for using too many short reading and listening passages, for testing students on non-academic content and for including short items that test low-level fluency.
Regular readers don’t need me to tell them that the revised TOEFL will not include meaningfully integrated speaking and writing tasks. Regular readers also know that it will include shorter reading passages than those currently found on the DET, and that many of them will be about non-academic subjects. And, needless to say, the new TOEFL will include many short items that test low-level fluency.
Do not take this as a complaint. The revised TOEFL is a wonderful test. I mention these things only to highlight how different it is from the old TOEFL and how much it resembles the DET.
In some ways, this report can be read as a sort of coda for the old TOEFL program, before ETS gave up the ghost and determined that the best way forward was to remake the test into something that very closely resembles Duolingo’s product.
So it goes.
Apropos of various things, the Duolingo English Test team has published a charming new video highlighting their mission of “transforming testing for good.” It features Duolingo co-founder Luis von Ahn and others. I was happy to see DET creator Burr Settles make an appearance.
Duolingo is really good at this sort of thing. And it accounts for so much. Some people tend to miss that. Those who consume their output, though, are quite aware of it.
Is this video part of their HOELT charm offensive? Probably.
The UK Home Office has published a fifth request for information regarding the Home Office English Language Test (HOELT). This one is a shocker. It notes that “the Home Office is exploring a ‘Digital by Default’ service, with remote proctoring as the primary mode of delivery and physical test centres available where remote solutions are not feasible.”
This could explain the curiously low number of test centers mentioned in the fourth RFI, which is again listed as just 268 in 142 countries.
A Home Office choice to go with remote proctoring by default might favor a smaller test provider – like LANGUAGECERT, Duolingo or ETS – heretofore considered an underdog in the race to win the tender. All three of those providers are well known for offering robust remote-testing options to test takers around the world.
On the other hand, the IELTS partnership (widely considered a front-runner to win the HOELT tender) currently offers remote tests only in select markets, while Pearson (another favorite) pulled the plug on its remote options back in 2024 shortly after stories broke about widespread cheating on the at-home PTE Test.
Of course this doesn’t mean remote testing is a sure thing. But it is worth paying careful attention to the possibility.
I want to cycle back to that pre-print from Maggie McGehee and Daniel Isbell which I wrote about a few weeks ago. Take a quick look at the data in Table 1. In the period under review (Fall ‘22 to Fall ‘23) 100 undergraduate students submitted ELP scores to the University of Hawaiʻi at Mānoa. Here are the tests they used:
- DET: 61
- TOEFL: 20
- IELTS: 19
And limiting ourselves to Fall ‘23:
- DET: 30
- TOEFL: 5
- IELTS: 8
So 61% of students submitting a score in the whole period opted to send a Duolingo English Test score. DET’s share reached 70% in the final semester under review, while the TOEFL had just a 12% share at that time.
What I wouldn’t give for numbers from 2019.
This is just a single less-selective school in the second most beautiful state in the union, but it speaks to the challenge that the TOEFL team faces as it implements a major relaunch of its flagship product.
That challenge being how to respond to a young person who knows that all his friends were admitted to good schools using a DET score, and who also knows that the DET is just a fraction of the price.
I wouldn’t know what to say.
Regardless, one can almost suss out the potential future (bright) of the TOEFL program. Some things come to mind:
- Maybe US admissions are something of a lost cause, but there are a ton of jurisdictions that are reluctant to accept DET scores right now. Places like the big new receiving countries in Europe and Asia. Perhaps TOEFL could become the go-to “easy” test for these markets. I can envision it becoming something like the DET of the EU. I would mention Australia here, but the revisions have come too late to be accepted by the DHA.
- On the other hand, maybe the domestic market can be recaptured. The new TOEFL is hella streamlined. Development, administration and scoring will all be much cheaper than for the old TOEFL. Why not turn TOEFL into a $75 at-home test and instantly recapture much of the market share lost to DET? Heck, even knocking the price down to a cool $99.99 would probably do the trick. A test center version could be maintained at a hefty price tag for folks who prefer it. To put it in somewhat nefarious terms, ETS could charge test takers in the lucrative Chinese market $300 a piece and let everyone else test for $100.
- On the vestigial third hand, there are big players who greatly desire a bigger slice of the American pie. Pearson has gone to all the trouble of developing their own DET competitor (the PEET Test) which will launch next month. That test looks fantastic, but it will take an enormous amount of time, money and effort to gain widespread acceptance at institutions. It took Duolingo six years and a pandemic to gain a stable foothold. Why not just… well… I won’t say it. But you know what I mean. Everybody wins.
The cool thing is that all three of these avenues suggest a bright future for the TOEFL program. They just require boldness and fortitude.
The Duolingo English Test’s Zendesk was recently updated to include notes about all of the question types which appear on the test. Each type is given a separate page with a description (including timing), tips for managing the user interface, and general notes about how to respond/answer. There is also a separate page titled “How the Items are Scored.”
A nice nod to the transparency we’ve been discussing recently.