By Bijendra Nath Jain, Visiting Distinguished Professor of Computer Science, IIIT Delhi
The cancellation of the National Eligibility cum Entrance Test (NEET) exam after allegations of a paper leak is a wake-up call to improve a system that millions of aspiring students rely on. The recurring disruptions in our high-stakes examination apparatus are not merely administrative oversights, but symptoms of reliance on outdated systems to generate and distribute question papers. To permanently secure examinations of the magnitude of NEET and Joint Entrance Examination (JEE), we must overhaul how and when question papers are created and distributed.
The traditional method relies on multiple teams of subject experts preparing draft question papers, followed by independent teams vetting and finalising a single paper. To maintain absolute secrecy, these experts are isolated for days leading up to the test. When exams span multiple daily sessions or require translation into regional languages, many more teams of experts and translators must participate and remain isolated, multiplying the administrative burden and the risk of question paper leaks. No matter how many layers of security are applied to secure physical printing and distribution of a paper, a static document existing days in advance is inherently vulnerable.
I propose that the final question paper be assembled in electronic form using a randomisation algorithm applied to a massive, publicly available bank of curated questions and loaded into student terminals minutes before the scheduled examination.
To ensure statistical integrity and prevent predictability, the question bank must be exceptionally large. If a paper must have N questions, the size of the bank should be 200 to 1,000 times N.
Sourcing that many is best achieved by crowd-sourcing potential questions, deploying multiple levels of expert committees to screen submissions, pre-test them for completeness and accuracy, remove ambiguities, and categorise them by subject, topic, and difficulty. Sourcing items in this way does not compromise secrecy since ultimately the entire question bank will be made public. More importantly, no human will have access to the finally compiled question paper that wholly relies upon a secret key used to randomly select questions from the bank.
This architecture supports a highly progressive pedagogical shift: making the entire question bank public for students and teachers. Providing open access to the repository won’t undermine the exam’s capacity to discriminate candidates based on their understanding. While safeguards must be developed to prevent rote learning, open access will facilitate testing (viz. curation) of questions in the bank and students to prepare using mock tests.
On the day of the exam, and minutes before the scheduled start time, the question paper is generated centrally using a randomisation algorithm initialised by a secret key, itself a random number generated just-in-time. Instructions to compile the question paper are distributed securely over a VPN directly to local servers in exam centres. To transition from the current system to a fail-safe digital ecosystem, the ministry concerned and the National Testing Agency (NTA) should commit to an implementation road map over the next 18 months:
Months 1-6: Question bank & infrastructure preparation: The NTA, or another agency, should immediately initiate crowd-sourcing of questions to start building the large question bank repository. Simultaneously, it should upgrade digital infrastructure across physical test centres to establish a secure interconnected set of servers capable of receiving encrypted files over a VPN.
Months 7-12: Curation, translation, and public release: Multi-tiered committees of domain experts must review, pre-test, translate, and categorise the submitted questions into distinct baskets based on subject, topic, and level of difficulty. Once created and tested, the entire bank of curated questions should be made public.
Months 13-18: Pilot execution and nationwide integration: The centralised randomization algorithm, secure network, and just-in-time distribution must be piloted extensively across India using localised or nationwide screening tests. Following validation, manual paper setting should be discontinued, shifting all candidates to touch-sensitive tablet devices evaluated instantly by computers.
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.
