At-Home or Test Centre? What Cambridge's Paper on Remote English Tests Means for You

A 2026 paper from Cambridge University Press & Assessment, Remotely Delivered Language Assessment: What Makes It Fit for Purpose?, argues that at-home English tests are convenient and widen access, but should be used for high-stakes decisions such as visas and university admission only when their security, fairness and score comparability match in-centre tests. The authors set out six principles (testing what matters, rigorous security, standardised conditions, human involvement, fair and inclusive participation, and comparability across modes) and say that for the highest-stakes uses "in many cases, in-centre delivery may be more appropriate than remote". For students the practical rule is simpler: choose the format the organisation you are applying to accepts, then choose the one in which you can perform at your best on the day. This guide summarises the paper's arguments and gives a checklist for choosing between a test centre and an at-home test.

What the Cambridge paper is and why it matters now

Remote testing (also called online or at-home testing) means sitting an exam on your own computer, in a place you choose, while being watched by a webcam, software or a remote human proctor. The paper notes that remote speaking tests by telephone existed in the 1990s, but the Covid-19 pandemic turned remote delivery into "a standard delivery option for assessments alongside in-centre delivery".

The question the authors ask is not whether remote testing is good or bad. They write: "This is not an argument against remote testing, but one for a more rigorous assessment of how, when, and in which contexts it can be used with confidence." The timing matters because governments are making very different choices. The UK Home Office is procuring a fully remote visa English test (see our explainer on the UK's remote English test plan), while the BBC has reported that Australia banned remote or at-home tests for migration purposes in August 2025.

ICEF Monitor summarised the paper as cautioning "against at-home English tests being accepted as part of students' visa application until sufficient measures are in place to ensure test security". Dr Evelina Galaczi, Cambridge's director of research for English, is quoted there: "When test results determine high-stakes decisions, the bar for security must be exceptionally high. Our research shows that while technology can help, it cannot eliminate risk alone."

Benefits and risks of at-home tests, according to the paper

Benefits the paper lists	Challenges the paper lists
Convenience and flexibility: no travel, more choice of dates, and a familiar room may reduce anxiety	Test security: less direct control than in a centre, so malpractice is harder to detect
Wider access: people in remote areas, or with some disabilities or mobility difficulties, can test more easily	Bias: some AI proctoring tools have raised concerns about racial or ethnic bias and may wrongly flag normal behaviour
Live speaking still possible: video-call speaking tests keep real human interaction	Proctor training: remote proctors, often outsourced, must be vetted, trained and audited
Lower indirect costs: less spent on travel	Privacy: ID documents and video of your home are collected and stored
	Technology and space: not everyone has a good device, stable internet or a quiet private room

Source: Lee, Isbell and Kremmel, "Benefits and challenges of remotely delivered language assessment", read 12 November 2026.

The paper also describes how cheating has changed. Copying a neighbour's answers is now hard because of computer delivery and randomised questions, but remote testing brings new risks: recording and sharing questions ("item harvesting"), using unauthorised resources, and impersonation, including "an imposter inside or outside the room providing answers via wireless devices". It adds that generative AI has made "highly convincing deepfake audio or video" more feasible, which could undermine remote identity checks.

The six principles in plain English

Principle	What it means	Question to ask a test provider
1. Testing what matters	The remote version must still test the skills the decision needs. A short, machine-prompted speaking task may not show whether you can hold a real conversation.	Does the speaking test include live interaction, or only recorded answers?
2. Rigorous security	Remote tests need stronger ID checks, monitoring and protections because they are more vulnerable to malpractice.	Is proctoring live, recorded and reviewed, or automated only?
3. Standardised conditions	Every test taker should get similar conditions, even though rooms and equipment differ.	What exactly must my room, desk and computer look like?
4. Human involvement	AI can help, but "cannot replace human judgement altogether", especially in high-stakes and regulated contexts.	Who reviews a flag before my result is cancelled?
5. Fair and inclusive participation	Remote delivery must work for people with different devices, internet, homes and disabilities.	What adjustments are available, and can I take the test in a centre instead?
6. Comparability across modes	If remote and centre versions are presented as equivalent, the scores must be shown, and kept, comparable.	What evidence shows that home and centre scores mean the same thing?

The paper adds a useful point about cost: "At a minimum, remotely delivered assessments should not be more expensive than those offered at test centres."

How remote proctoring works: six levels of security

The paper includes a "hierarchy of remote proctoring options", adapted from earlier research by Michel (2020). From most to least secure:

1. Live proctoring with AI technology: a person watches in real time, supported by software alerts.
2. Live remote proctoring: a person watches in real time.
3. Record and review with AI technology: the session is recorded, software flags moments, a person reviews them.
4. Record and review later: the session is recorded and reviewed after the test.
5. No interaction, AI only: software monitors with no human involvement.
6. No interaction, honour code: the test taker promises not to cheat.

The authors say the right level depends on the stakes: "In high-stakes cases, e.g., for certification or admissions purposes, the most secure level of proctoring is needed, and in many cases, in-centre delivery may be more appropriate than remote." They also describe post-test checks such as statistical analysis of answer patterns and face or voice matching as "a crucial part of a layered security approach, as no single proctoring configuration can prevent or detect all forms of misconduct".

Reading the paper critically

The paper is careful and well referenced, and it openly lists the benefits of remote testing. Still, three things are worth keeping in mind:

• It comes from a test provider. Cambridge offers both remote tests (the paper names Linguaskill and IELTS Online) and in-centre tests, and is one of the IELTS partners that declined to bid for the UK's remote visa test contract. Its view on where remote testing is "fit for purpose" is informed but not independent.
• It is a position paper, not a new study. It sets out principles and examples. It does not publish new data on how often people cheat in remote versus centre tests.
• The coverage you may read is sponsored. The ICEF Monitor article was published in partnership with Cambridge English, so it presents the paper's arguments rather than testing them.

None of this makes the principles wrong. The six questions in the table above are reasonable to ask of any test provider, including Cambridge.

How to choose between a test centre and an at-home test

For most students, the decision comes down to two questions, in this order.

1. Will the organisation accept it? This is the only question that can rule a format out. Check the exact test name and delivery mode on the official page of the university, employer or government you are applying to. For example, the UK accepts only the tests on the Home Office SELT list for visas, and these are taken at approved centres; the Duolingo English Test and other online tests are not on that list (see our guide to approved SELT tests). Universities often have their own, wider lists.

2. Where will you perform better? If both formats are accepted, compare them honestly:

Choose a test centre if…	Choose an at-home test if…
Your home is noisy or shared, or your internet drops	You have a quiet, private room and reliable internet
You want staff to handle technical problems	You are confident setting up a camera, microphone and secure browser
You worry about being flagged for looking away or moving	Travel to a centre would be long, costly or stressful
You need adjustments, such as a scribe, that may only be offered in a centre	A familiar setting lowers your anxiety
You need the result for a visa or another decision where only centre tests are accepted	You need a date sooner than the nearest centre can offer

If you are still unsure, take a full timed practice test under the conditions you would have at home: same room, same time of day, no notes, phone in another room. If it goes badly for reasons that have nothing to do with your English, book a centre.

Practical checklist for an at-home test

1. Read the provider's room rules the week before: what can be on the desk, whether you can wear headphones, whether you may look away from the screen.
2. Run the system check on the same computer and network you will use on the day.
3. Clear the room. Remove notes, extra screens and devices. Tell people at home not to enter.
4. Prepare valid ID that exactly matches your booking name.
5. Practise speaking to a screen. Many remote speaking sections use recorded prompts with a timer. Our free Speaking Lab helps you practise answering out loud.
6. Keep to the rules even when it feels silly. Reading questions aloud or looking away to think can trigger a flag on some systems. Ask the provider in advance what is allowed.

Whatever the format, the preparation that raises your score is the same: daily speaking, listening at real speed (see how to improve English listening) and timed writing with feedback. Our exam preparation courses and free lessons cover B1 to C1.

What this article cannot tell you

• Which tests a specific organisation accepts. Lists differ between countries, universities and visa routes, and change often. Check the official page before you book.
• Real cheating rates. We found no independent published data comparing malpractice in at-home and in-centre English tests. The paper argues from principles and examples, not new statistics.
• Every detail of the paper. We read the summary, the six principles, the benefits and challenges, and the security sections. Later sections give examples from Cambridge's own tests, which we have not evaluated.
• Prices. Fees vary by test and country, so we have not listed them.

This page is independent guidance from Oxford English Global. It contains no affiliate or sponsored links.