Are the standardized tests good measure of academic ability or progress?

The Thermometer and the Fever: What Standardised Tests Can and Cannot Tell

Picture two seventeen-year-olds in the same city. One wakes at five to attend a coaching institute, solves thousands of practice questions and learns to finish a three-hour paper with minutes to spare. The other, without such help, reads widely, builds a small solar lamp for her village and asks questions her teachers cannot answer. On the morning of a national entrance test, a single score will decide which of them is "able". The score is not meaningless; it measures something real. Standardised tests are a good measure of a narrow band of academic ability at a single moment, and a poor measure of the whole of ability or of a learner's progress, unless they are read alongside richer evidence of how a child thinks and grows.

It helps to be clear about the terms. A standardised test asks every candidate the same questions, under the same conditions, scored in the same way, so that results can be compared across lakhs of people. Such tests serve two very different purposes. Some are high-stakes selection tests, such as JEE, NEET, CUET, China's gaokao or the American SAT, which decide individual futures. Others are low-stakes diagnostic surveys, such as the National Achievement Survey, the Annual Status of Education Report or the OECD's PISA, which judge systems rather than children. Equally, academic ability is a person's capacity to learn and reason, while progress is the change in that capacity over time. A test can capture one far more easily than the other.

The case for standardised tests is strong, and it begins with fairness. Imperial China used written civil service examinations for centuries precisely to replace birth and patronage with merit. India's own civil services examination is admired because a farmer's daughter and a minister's son sit the same paper and are judged by an anonymous examiner. Uniform questions shield candidates from the bias of a local teacher or the influence of a powerful family. Moreover, tests scale: no other method can fairly rank twenty lakh medical aspirants in a matter of weeks. When several American universities dropped test requirements during the pandemic, the Massachusetts Institute of Technology reinstated them in 2022, explaining that scores actually helped it identify talented students from under-resourced schools whose other credentials were thin.

Standardised measurement is also indispensable for seeing progress at the level of a system. For years, India celebrated near-universal enrolment while children learnt little; it took the household tests of the Annual Status of Education Report to show that large numbers of Class 5 children could not read a Class 2 text. When India took part in PISA in 2009 through Tamil Nadu and Himachal Pradesh, the two states ranked near the bottom of the table, a humbling verdict that no internal examination had delivered. Germany's own "PISA shock" of 2001 triggered a decade of school reform. In this role, a test works like a thermometer: it does not cure the fever, but without it we would not know the patient was ill.

Yet a thermometer measures only temperature. Standardised tests measure what is easy to measure: recall, speed and the ability to choose among options under pressure. The psychologist Howard Gardner argued in 1983 that intelligence is plural, spanning linguistic, logical, spatial, musical, bodily, interpersonal and intrapersonal capacities. A multiple-choice paper sees two of these at best. Creativity, curiosity, collaboration, moral courage and the patience to work on a problem for months are invisible to it. A single morning's performance is also hostage to illness, anxiety and luck. Many of history's great minds were indifferent examinees, and many toppers never produced an original idea.

Building upon this, high stakes corrupt the measure itself. The social scientist Donald Campbell observed that the more a quantitative indicator is used for social decision-making, the more it will be distorted, an idea close to Goodhart's law. Schools begin to teach to the test, and students learn to game formats rather than master subjects. The coaching economy of Kota, with its "dummy schools" where students are enrolled but rarely attend, is the logical endpoint. Scores then reflect access to coaching as much as ability. The Justice A.K. Rajan Committee, set up by Tamil Nadu in 2021, concluded that NEET had favoured students from affluent, English-medium and coached backgrounds. The human cost appears in the distress of adolescents for whom one number has become the measure of their worth.

A test is also only as good as its integrity. The NEET-UG controversy of 2024, with allegations of paper leaks, shook the trust of millions of families, prompting the Centre to set up a high-level committee under former ISRO chairman K. Radhakrishnan to reform the National Testing Agency, and to bring into force the Public Examinations (Prevention of Unfair Means) Act, 2024, which Parliament had passed earlier that year. More fundamentally, a one-time test cannot capture progress, because progress is a trajectory, not a point. That needs formative assessment: regular, low-stakes checks by teachers who know the child. India's attempt at Continuous and Comprehensive Evaluation under the Right to Education Act was well-intentioned but poorly implemented, and the 2019 amendment allowing states to end the no-detention policy showed how hard it is to measure growth without slipping back into exams.

Other nations have struck different balances. Finland, long a top performer, has no national standardised tests in basic education apart from sample-based evaluations, and relies on highly trained teachers to assess pupils until the matriculation examination at the end of upper secondary school. China, the home of the gaokao, launched its "double reduction" policy in 2021 to curb homework and for-profit tutoring after recognising the toll of test pressure. India's National Education Policy 2020 charts a middle path: board examinations that test core competencies rather than rote memory, a 360-degree holistic progress card that records growth in many domains, and PARAKH, a national assessment centre to set standards for assessment.

The synthesis, then, is not to abolish standardised tests but to put them in their place. For selection, they should be one input among several, alongside school records, aptitude for the field and, where feasible, interviews or portfolios. For systems, low-stakes sample surveys give governments honest data without punishing individual children. For learners, teachers' continuous observation, projects and self-assessment trace the progress that a single score misses. Better test design, from adaptive testing to competency-based questions, can also narrow the gap between what is tested and what matters.

Return, finally, to the two students. A good standardised test can tell us, fairly and cheaply, that one of them solves physics problems faster today. It cannot tell us which of them will invent something, lead a team or keep learning at forty. Tests are an honest thermometer of a narrow kind of ability and a useful census of a system's health, but they are a poor portrait of a person and a weak record of growth. A wise education system will use them for what they do well, refuse to let them decide what they cannot see, and remember that the purpose of measuring learning is to improve it, not to rank children for its own sake.