Can adaptive testing break the coaching cartel? Rethinking JEE Advanced and the future of fair merit

India’s competitive-examination ecosystem for admissions to premier institutions such as the IITs and the IIMs has spawned a coaching industry excelling in training students to ace the exams without always expanding their knowledge base as much. Against this backdrop, the IIT Council’s recent proposal to explore adaptive testing for JEE Advanced can be a break-through reform in Indian education system. This article studies the current testing system, advantages of adaptive testing, track record of global testing agencies, challenges in the introduction of the same in India and suggests a way forward.

The multi-billion dollar test preparation industry

As per market research firm Imarc, India’s coaching industry, estimated to be over ₹58,000 crore ($6.5 billion) in 2024, operates under a largely unregulated environment, with enrolments of over 7.1 crore students. It is projected to grow at a CAGR of over 10.4% to hit $17.4 billion by 2033, trapping millions of aspiring students, with the promise of success.

The intense pressure to succeed amidst stiff competition has led to a growing mental health crisis, especially in coaching hubs like Kota, where the growing number of student suicides has drawn national attention. For the parents, it has turned out to be a massive financial drain, with annual coaching fees ranging from ₹80,000 to ₹2 lakh, forcing some middle-class families even to sell their properties or take high-interest loans, with the hope of securing a bright future for their children.

Parliamentary panel to probe on coaching centres

On November 25, 2025, the Lok Sabha formally announced, through bulletin no. 2505 constitution of the Standing Committee for a comprehensive probe into the proliferation of coaching centers. The aspects of investigation include the “Dummy School” phenomenon, where students skip regular schooling to attend coaching classes; the alarming student suicide rates; and also whether the current competitive tests give an unfair advantage to the students who can afford these expensive coaching classes.

Adaptive testing for JEE Advanced

As per the minutes of the last IIT Council meeting released recently, the council recommended a phased shift to AI-driven adaptive testing for the JEE Advanced, to minimise the influence of the coaching industry and reduce the stress on students. The council noted that the current pattern of linear examinations rewards rote memorisation and pattern-matching, practiced in the coaching centers. The council proposed transition to Computerised Adaptive Testing (CAT), which allows the system to evaluate critical thinking and reasoning. By dynamically adjusting difficulty, the test can more accurately identify high-ability candidates without requiring an exhausting fixed-length test paper.

Traditional testing methodology

In a traditional competitive test, every candidate answers the same set of questions with average difficulty across the test paper. Test scores are based largely on the number of correct answers. Most of the coaching centers prioritise training on test cracking tactics over conceptual understanding, critical thinking, or problem solving. Students are trained through repetitive mock tests designed on the basis of previous question papers so as to simulate the pattern of the real examination. By practicing repeatedly, the students develop a reflexive speed that may mask their true conceptual understanding and perform better than other students.

What is adaptive testing and how is it different from traditional testing?

Adaptive testing is a dynamic form of assessment which uses the Item Response Theory (IRT). This means that the questions in the test are not the same for all the students. A question bank is prepared with questions of varying difficulty covering all the items in the syllabus. The test normally begins with a question of medium difficulty. If it is answered correctly, the computer selects a more difficult question from the question bank using the scientifically established psychometric models. If incorrect, an easier question is presented. After each answer the computer recalculates the candidate’s estimated ability. The test ends when a pre-defined level of assessment is reached or the set number of items as per syllabus are tested. The goal is to assess the ability of the candidate more precisely by asking fewer but better-targeted questions.

The result is that two candidates may see different questions but are assessed on the same underlying scale. This is fundamentally different from multi-shift or normalised exams where different papers are statistically adjusted after the test.

Adaptive testing builds fairness into the test design itself rather than correcting for variations subsequently. Since the next question depends on the previous answer, unless the student is conceptually sound, he/she can not answer correctly the progressively higher-difficulty questions, which have higher weightage.

Experience of global competitive tests

Global testing agencies have been using computerised adaptive testing for admissions to higher studies in premier international institutions for over 25 years. Over a period of time, they have refined their methodologies on the basis of experiential feedback from the students.

Graduate Record Examination (GRE), used for admissions to Post-graduate programs in Engineering, Management and Law, is taken by over 250,000 students per year. It started using item-level adaptivity in the 1990s and in 2011, switched over to Multi Stage Testing (MST), at the section level.

If a student does well in one section like quantitative section, the questions in the other sections, like logical ability are more difficult. In order to address the issue of the anxiety of “one-question-at-a-time,” MST allowed the students to see all questions in a section and navigate freely within the section which significantly improves the test-taker experience.

The Graduate Management Admission Test (GMAT), an entrance examination conducted by the Graduate Management Admission Council (GMAC) is used for admission to business schools globally for MBA programmes and is taken by over 115,000 students per year. It transitioned to Computerised Adaptive Testing (CAT) in 1997. In order to facilitate back-tracking and skipping of questions, it allows students to bookmark and edit up to three answers per section, balancing adaptivity and retaining control by the students.

The experience of GRE and GMAT competitive examinations demonstrate not only the benefits of adaptive testing but also the scalability at an international scale provided the governance and psychometric foundations are strong.

Indian experience

In India, NMAT (Narsee Monjee Management Aptitude Test), conducted by GMAC for admission to the MBA programme of Narsee Monjee Institute of Management Studies (NMIMS) and accepted by some other business schools, has been using the adaptive testing model since 2015. Over 75,000 students are reported to have taken the test last year.

It is widely believed that NMAT has enabled more personalised assessment of candidates enhancing fairness and ability assessment. Students are allowed flexible scheduling within a 75-day test window and the ability to choose section order and multiple retakes (up to three per year), improving convenience and individual comfort. Transition challenges from the earlier system were addressed through clear communication and guidance to the students including practice resources like official mock tests.

Legal challenges in Indian environment

One of the biggest hurdles expected by the IIT Council and the National Testing Agency (NTA) is the potential for legal challenges. In the Indian constitutional context, ‘equality of opportunity’ is often interpreted as everyone getting the same paper. Legal experts suggest that while the questions differ, the standard of measurement is identical.

However, the move may face Article 14 challenges (Right to Equality) in Indian courts, if the normalisation process is not transparent. Algorithm opacity invites bias suits under equity mandates; DPDP Act scrutiny on data; appeals over “black-box” ranks. However, it can be mitigated through disclosures and equity audits.

Transparency in a high-stakes culture

In India, where a single rank can decide a rewarding career, the transparency gap is a major challenge. It may be difficult to convince a parent how a student that solved 20 difficult questions correctly has scored higher than another student who solved 50 easy ones. Ensuring availability of lag-free technical infrastructure in tier-3 cities is also a challenge as any technical glitch during the test could be seen as maladministration in a court of law. Without clear documentation and grievance mechanisms, litigation risk rises. Data centre reliability, proctoring invigilation systems, and incident handling must be far stronger than today’s baseline

Transitioning from traditional linear test to adaptive tests requires building large calibrated item question banks, pretesting questions without possibility of leakage and ensuring strict syllabus coverage. There are also trust challenges: candidates must believe that different questions can still be fair and the testing agencies must be prepared to defend this scientifically and legally. International experience from GRE and GMAT shows that transparency, technical documentation, and phased rollouts are essential for acceptance.

IIT council’s proposed roadmap

The IIT Council’s transition plan aims to neutralise the deleterious influence of coaching through a phased approach spanning over two year period (2026 to 2028). It includes the launch of free and optional adaptive mock tests in 2026 to calibrate the item bank and familiarise students with the interface. It proposes to move to section-level adaptivity in 2027, followed by full integration with IRT-based scoring, moving the national merit list from “Raw Marks” to “Ability Scores.”

Can JEE Advanced become the template for other competitive examinations ?

JEE Advanced is a logical starting point because of its academic ownership, controlled scale, and question-paper-setting diligence. A successful pilot will provide a blueprint for other competitive exams and even recruitment testing, where adaptive screening could improve efficiency and fairness. However, each use case would require strong accessibility provisions, legal defensibility, and transparent score interpretation.

Can the adaptive testing threaten the coaching market?

India cannot eliminate coaching entirely, as any high-stakes selection system creates a potential business for test preparation. But adaptive testing can avoid the trends of rote drilling and pattern memorisation and shift training toward conceptual understanding and critical thinking. If the IIT Council’s JEE Advanced proposal is implemented with psychometric rigour and transparency, it could mark the beginning of a radical transition from pattern-based tests to ability measurement-based selection.

Conclusion: The way ahead

If the proposed plan by the IIT Council goes through, by 2028, we can expect a unified national adaptive testing framework, which may become the norm for competitive tests like JEE. The pedagogical landscape can shift from rote memory and shortcuts to improving conceptual clarity and problem solving. It will ensure more fairness in selection, on the basis of merit alone, enhancing inclusivity and quality of education. It will be a major step in the journey of transformation of Indian education, in line with the objectives of NEP-2020.

(Prof O.R.S. Rao is the Chancellor of the ICFAI University, Sikkim. Views are personal.)

(Sign up for THEdge, The Hindu’s weekly education newsletter.)