Re Education : Issue #31 - The Examined Life
Are exams the worst form of assessment...apart from all the others ?
(If this message does not format properly in your email - please click the “Read in App / View in Browser” above, top right).
Quote of the Month
The unexamined life is not worth living
Socrates
Introduction
Socrates was undoubtedly thinking of self-reflection and introspection, not public high-stakes examinations, when he is said to have uttered his famous dictum.
The lives, or rather learning, of young people in modern society are rather more examined than not, often in extremely narrow ways. High-stakes exams, for entrance to secondary or higher education, can be life-changing, for both good or ill.
Yet curiously, given their importance, and the impact examinations have on teachers and the curriculum, very few education development programmes in the global south include examinations reform as part of their support to a Ministry of Education – why not ?
Andy Brock, May 2026
TURN OVER YOUR PAPER NOW
Q1 : Do Exams Work ?
One reason may be the overwhelming focus of development support, at least in the last 10 years, on basic education. Perhaps understandable given the “learning crisis”, but shortsighted in terms of supporting national development and creating the social signals that encourage learners with the promise of better economic opportunities.
That promise feels shakier than ever because it’s consistently undermined by the gatekeeping function of national exams. For example, in Kenya, only 27.4% of students achieved the C+ grade required for university entry in the 2024 KCSE and in Nigeria, the 2025 WASSCE saw only 38.32% of nearly two million candidates qualify for university admission.
Such figures serve to emphasise the overwhelmingly academic focus of secondary education transitions, yet many question whether these exams are actually measuring learning, even for university. Parthajeet Das of Central Square Foundation, an Indian NGO, says :
Our examinations are still testing for memory, testing for what you have crammed in your head…as opposed to testing for learning
Lawrence Dogbatse, Deputy Register of the West African Examinations Council (WAEC), Ghana, agrees :
Generally, examinations, especially in Ghana and West Africa, still do a reasonable job of measuring content knowledge, recall and the basic understanding. However, often they fall short of assessing competencies, critical thinking, creativity, collaboration and practical skills.
In 1973, Kenneth Arrow, an economist, wrote a seminal paper called “Higher Education as a Filter” – arguing that education qualifications (and the exams that lead to them) acted as a screen for higher education and employment alike, testing for performance ability rather than acquired skills. Little seems to have changed since then.
A recent book about the “gaokao”, the Chinese university entrance test, called “The Highest Exam”, (Ruixue Jia and Hongbin Li – see my review below) points out that there is evidence to show that Chinese students’ learning stagnates at university, reinforcing the notion that high-stakes exams are gates to be hurdled not summative assessments of development.
All this might just be acceptable if there was a level playing field in test taking – but, this is not the case. Social class and family circumstances tend to be among the best predictors of success in exams : nearly 30% of the places in top UK universities go to the 7% who are privately educated.
Georgie Billings, Head of Assessment Quality, Cambridge University Press and Assessment, says exams :
…tend to encourage negative washback and over reliance on drilling. They also tend to privilege students who can afford to have private tuition.
Parthajeet Das adds that the language of exams seldom accommodates the mother tongues of many Indian communities, and notes that in India another equity issue is a lack of standardisation of examinations between different states. Jia and Li point to a further aspect – the persistent rural / urban bias in the gaokao, a phenomenon widely found in other countries.
Adding to this litany of criticism, is the uneven provision for children with special needs. In Public Examinations Examined, written pre-Covid, the authors say :
While systems in which participation rates are low may not yet have come to terms with these issues, they should be prepared to deal with them as participation rates increase.
Q2 : Assess the Washback Effects of Examinations
So much for pupils, what of the impact on teachers and the curriculum of high-stakes examinations ?
Research into the impact of high-stakes examinations, where teachers face accountability pressure tied to results, shows that teaching to the test is not an aberration, rather it’s a rational response – sometimes by individual teachers, sometimes by whole schools, or even education systems.
The higher the stakes, the earlier teaching to the test starts and the narrower the range of learning covered, the more the content conforms to what can be measured and tested. In “The Highest Exam” Jia recounts how in high school the first two years were exclusively devoted to the content of the gaokao and the third year almost exclusively to exam preparation with weekly testing.
Unsurprisingly, such pressure create two additional temptations – cheating and external coaching.
A study in Indonesia of widespread cheating on national exams showed that when the government introduced computer-based testing between 2015-2019 (in a bid to prevent cheating) test scores dropped dramatically by about 0.5 standard deviations. But, over the next two years, scores rebounded showing that teachers and learners had to shift their focus onto actual learning. Not all washback is bad.
In some countries cheating spills into coaching and the creation of an entire shadow tutoring system aimed at passing exams. A paper by Ian Hardy et al published in 2025 on the effect of the “shadow education” system in Bangladesh argues that such systems go beyond a response to high-stakes pressures into a kind of institutionalised cheating – one only available to those with the resources to pay.
Mark Bray’s 2021 analysis of private supplementary tutoring in Africa also found extensive evidence of exam-led teaching and its social and cultural foundations across the continent — with specific documented examples from Botswana, Ghana, Malawi, and Tanzania.
But, perhaps the greatest washback effect is the perpetuation of the widely held belief that exams are the culmination of a meritocratic process in which all have equal opportunity. Georgie Billings says :
It’s a kind of meritocracy lie really, the idea that exams are some sort of great social leveller.
Sammy Wright in his recent book “Exam Nation”, about the English education system (see my review below) goes further :
..the heart of the problem is that the idealised “meritocratic” logic of the system results in schools whose purpose is not the general education of all pupils according to their natural aptitudes, but rather the sorting of kids into the positions in society that they “merit” according to an academic scale. The most pernicious result of this is that the pupils themselves can see the lie for what it is.
But, not all education systems are dominated by examinations. Some of the best performing systems in the world place very little emphasis on high-stakes exams. In Finland, Iceland, and Sweden, the upper secondary diploma that students receive is based entirely on teacher-assigned grades. Norway, Denmark and New Zealand also have large elements of the final assessment dependent on teacher assigned grades. Pressure has not been removed, rather it has been distributed across years of continuous work and assessment.
Q3 : Why is it so difficult to reform examinations ?
Given the evident shortcomings of examinations, and the fact that working alternatives exist, why are high-stakes examinations so difficult to reform ?
Moreover, why do donors and philanthropists studiously avoid supporting, or even advocating for exam reform, when it is probably the single area most likely to create impact, especially for disadvantaged children, in an education system ?
Rachel Outhred, Managing Director of Oxford MeasurEd, is direct :
I think the major reason why it’s difficult to reform exams is that the systems are working exactly the way that they were designed to work. And those who have the power to reform exam systems are the ones that currently benefit from it the way that it currently is.
Jia and Li point to a linked reason – the power of parents (largely middle class), even in an autocracy like China. Li worked on a project to reform the gaokao for over two years until it was suddenly cancelled without explanation. He reflects :
…China’s people have relied on one test for their whole lives. They trust the gaokao. Reforming the gaokao is the equivalent of reforming the entire education system.
In this respect parents are simply responding rationally to the rules of the game as they change. In Kenya a major initiative to abolish the Kenya Certificate of Primary Education (KCPE) in order to reduce the scramble for placement slots at secondary school may already be running into trouble as parents work out how to gain competitive advantage in the new system.
This neatly demonstrates the equity paradox of examinations – they are simultaneously a source of inequality and the instrument that prevents worse (but usually different) inequality. They are, to paraphrase Churchill, the worst form of assessment…apart from all the others.
Even where serious attempts have been made to reform high-stakes examinations progress has been difficult. In Singapore attempts to shift from academic stratification to holistic development over the last two decades have simply led to more parental investment in tutoring, as the terrain for gaining competitive advantage has started to shift.
Georgie Billings points out that real reform of examinations into more competency-based assessments “requires cultural acceptability of those forms of assessment as valid and reliable” in order for it to be widespread. Additionally, the investment in teacher training needed to change from examinations based to continuous assessment is a huge barrier to change.
Lawrence Dogbatse agrees and points out that, as Ghana has moved to make 30% of the West African Senior School Certificate Examination (WASSCE) assessment based, the training of teachers and the formation of Professional Learning Communities (PLCs) for teachers has required a significant investment of time and resource. Continued development of continuous assessment, he says, will be possible only if teacher integrity and quality assurance are strong.
Perhaps reform will be driven by the feet of parents not bureaucrats. In India some private and not‑for‑profit schools are experimenting with broader assessments, portfolios, and holistic development, often in response to parental demand. Parthajeet Das says parents are increasingly sceptical of narrow exam success and recognise that exam scores alone do not guarantee access to higher education or employment, especially as many universities rely on their own entrance tests.
Q4 – Will Artificial Intelligence (AI) change examinations ?
Given all the barriers to change that stand in the way of reforming examinations, it’s tempting to see AI as a potentially significant disruptor.
Proponents argue that the individualised learning support which AI will increasingly offer has the potential to be matched by individualised continuous assessment – potentially lessening the additional workload for teachers. Others point out that, as AI improves, the fear of subjective judgements that bedevil non-exam-based assessment may recede.
In fact, that fear even applies to exam-based assessment as Daisy Christodolou in her piece “The Ethics of AI Assessment”, notes. She reports that with human marking, in GCSE English Literature, you currently only have a 52% chance of getting your true grade.
Parthajeet Das for one is cautiously optimistic. The key thing, he says, is that the promise of AI in assessment is that “it solves for scale”. Rachel Outhred also sees the potential in AI to help with assessment but makes a critical distinction : “I think AI will help us with examination products, but not with examination systems.” It may increase efficiency and standardisation, she says, but examination systems are too complex and too political.
Lawrence Dogbatse provides a note of scepticism. The value of immediate feedback that AI promises is clear, he says, but : “what’s it trained on and is it replicating colonial biases ?”
Georgie Billings is less certain that AI assessment is the direction of travel. She reflects that for AI to really play a significant role in assessment there needs to be a high degree of public trust. She points to an upsurge of interest in viva or oral examinations because they are hard to game and she wonders instead whether AI may better play a role alongside portfolio assessment, deepening and enriching holistic assessment in ways we can’t yet envisage.
ONE MINUTE LEFT
Perhaps, given the complexity and the highly political nature of examinations reform it’s little wonder donors and philanthropists have been leery of entering this space. Yet, as the political economy movement within the sphere of education development gathers momentum, can it be avoided much longer ?
The tension between AI personalised learning and high-stakes exams is set to become the defining characteristic of education development over the next decade – and one where external supporters of education development in the global south may have a role to play.
High-stakes exams won’t soon be replaced by AI due to the signalling and filtering functions they serve. But, in addressing this tension there is an opportunity to reshape assessment in education systems more radically, more holistically as a result of the likely disruption to the status quo that AI is going to cause.
Nonetheless, any movement away from the comfort blanket that exams represent will require widespread debate and extensive public awareness in order to create trust and confidence in alternative models.
References and Resources
(Research supported by Claude)
Higher Education as a Filter Kenneth Arrow (1973)
Shadow Education in Africa: Private Supplementary Tutoring and its Policy Implications Mark Bray (2021)
Testing times: The affective impact of tests in Bangladeshi schools Ian Hardy, M. Obaidul Hamid, Vicente Reyes, Louise G. Phillips (2025)
Assessment Reforms in Singapore Tan, K (2022)
Public Examinations Examined Kellaghan, Thomas; Greaney, Vincent M. (2019)
The Ethics of AI Assessment Daisy Christodolou (2026)
“The Highest Exam” Ruixue Jia and Hongbin Li (2025). My review of the book.
“Exam Nation” Sammy Wright (2025). My review of the book.
News
The world out of school population has risen for the seventh year in a row – standing now at 273m according to the World Bank.
Infinite Jaz, a newsletter written by a Jewish American, Jasper Diamond Nathaniel, about Israel-Palestine affairs, reports on a really disturbing new phenomenon : deliberate settler attacks on schools in the West Bank. As he puts it “What would happen if this occurred the other way around? If Palestinian militants from the West Bank broke into Israel or a Jewish settlement and shot up a school, killing a child and an adult and wounding four others? Is there any doubt that this would be breaking news in every major outlet of the Western world?”
According to UNICEF more than 200 children have been killed in Lebanon by the IDF since the 2nd March. Remarkably, even amidst all this inhumanity there are astonishing and inspiring stories, Ben Phillips posted this about the Palestine Marathon.
A new report on initial teacher education has been published by the World Bank : From Prospective to Prepared Teacher: Initial Teacher Education. The launch is on 2nd June.
Euan Wilmshurst’s indispensable guide, the Global Education Events Guide, has transitioned from spreadsheet to website. It’s free, shareable and editable. Also invaluable.
Development
Interesting piece by Richard Ough an ex-DFID Economist, called “Friendly Fire : How Aid’s Defenders Helped Its Critics”. He articulates something I’ve felt and been writing about sporadically for some time – how quickly many in the aid sector accepted the cuts, and rather than fighting back and defending aid with all its imperfections, leapt to get to the forefront of the new paradigm. Talking of the sector he says :
Most of all, it would have refused the premise that aid’s imperfections invalidate its purpose…Some old habits deserve to die. But there is a time to improve the ship and a time to stop it sinking.
While Ough has some good points, Dan Banik’s piece “The crossroads may be the best road” shows African development specialists struggling to see the future through the fog. A lot of the comments he cites have the feel of wishful thinking, hope without (yet) substance to feed it.
I came across an organisation recently who could potentially be of use to NGO’s and small charities in the education sector : A4ID. They are a UK-registered charity with a global network of legal firms who offer pro bono services to smaller organisations who may not be able to afford expensive legal services. They have a presence in Kenya, India and the UK and their support covers over 150 countries. Their website is www.a4id.org and this link shows a specific piece of legal work they have helped coordinate supporting girls’ education in Afghanistan.
You’ve indulged in cognitive offloading, now try “cognitive surrender”, as described by the Economist. What’s next - cognitive abdication ?
David McNair’s post identifies something long known in the development sector – most citizens haven’t a clue how much is spent on aid (usually overestimating by as much as 20x). Yet, politicians routinely lament the lack of support for development from the public, which is their job to explain. A straw man held up by straw men.
No country is immune to the digital curse. The Chinese state is trying to get people reading more – but, only the “right” kinds of books of course.
AI Spy
In “Pakistan’s AI Reckoning” Hassan Aslam Shad makes a pithy and well-argued case for Pakistan to quickly move to address structural lags in the approach to AI, including the observation that devolved responsibility for education requires urgent coordination. “What is missing is a mechanism that requires coordination rather than merely encouraging it. Without it, even well-designed initiatives risk working at cross purposes.”
Fab AI report the results of an eight-week RCT in Sierra Leone to understand how Gemini models impact student outcomes. Results indicated an improvement of roughly 1.2 to 1.7 years of typical learning progress in low-and-middle-income countries.
Voices from the front
The World Bank has approved a $600m loan to the Philippines to address learning gaps – covering over 21m children. Unusually for the Bank, all the money is to be spent on “software” i.e. no buildings.
Mark Smith has a pop at the Global Education Evidence Advisory Panel (GEEAP) asking “When is a bad buy not a bad buy”. A room full of books is different from a library with a trained librarian he argues. It’s a strong case.
The Ministry of Education in Chile becomes the latest in what is now a flood of countries and states banning mobile phones in the classroom.
Voices from the rear
(Gray and Published Research)
Malala Yousafzai has a new TED talk (11m) in support of girls in Afghanistan and across the world. “If we start with something, work together and stay ambitious, hope stops being something we wait to feel and becomes something we create”.
New research from Greer at al “Supporting children’s mental health and emotional wellbeing in and through education in contexts of conflict and protracted crisis: A narrative, thematic evidence review” synthesises more than 100 studies since 2020. Local contextualisation of response seems to be the key need.
UKFIET has published a new set of blogs from South Korea, India, Pakistan and Kenya. The most intriguing one is “The sound of equity: Teachers’ school journeys captured through soundscapes” in which Singal, Khanna and Diksha in their research with teachers with disabilities use soundscapes as a reflective tool for teachers with visual impairments. Shame they couldn’t have shared an example rather than only describe textually.
Open PEA, the brainchild of Peter Evans and Stefan Dercon, recently published a very readable and pithy paper called “Cash versus Maize in Malawi”, lead author Edward Archibald. This is Open PEA’s first full product – more is to come aiming firmly at the development mainstream, not the PEA niche.
What’s relevant for those of us interested in learners learning more than eating is the forensic analysis of the system, the system actors and the interplay of incentives that leads to poor outcomes for the country, and the propositions for political feasible change.
You could apply the same to many education systems – and in fact, hot off the press, there’s a new short paper on the politics of education PPPs in Nigeria by Open PEA’s Kene Nwagbo.
Podcasts, Films and Books
If you missed it earlier, I have reviewed two recent books about exams that I would highly recommend.
I found Top Class’s recent podcast fascinating : “How Estonia delivers world class education : a conversation with Estonia’s Education Minister”. The discussion is particularly engaging on Estonia’s response to AI. Kristina Kallas, the Minister, has a bold take on what it means for teachers – more of them ! It’s also clear that teachers in Estonia are highly respected – more powerful than politicians - and the country spends 7% of GDP on education. Estonia - the place to be !
And finally….
An astonishing chart published in the Financial Times 16th May showing a correlation between the take off of smart phones and falling fertility rates. Correlation is not causality (see photo below), but the chart is nonetheless pretty arresting.
If you know someone who would be interested in reading this newsletter, please pass on, by clicking the share button below. Subscription is free - subscribers receive each issue of Re Education automatically every month.