EducationOpen access · Free to read
Empirical Study on Cut-Off Variance in National Entrance Examinations (2015–2025)
By Dr. V. R. Kulkarni & Research Cell · Published 26 July 2026 · 8 min read
Comprehensive analysis of 1.2 million national entrance exam scores revealing a 42 percentile gap in cutoff criteria between unreserved and quota tiers.
Abstract
This 10-year empirical audit analyzes test performance data from 2015 to 2025 across NEET, JEE and State Public Service Commission exams. The findings demonstrate that cut-off score disparities have expanded significantly, while rural public school students remain underrepresented across all categories due to baseline primary education deficits.
Key Findings
- ✓Cutoff gap spread averaged 42 percentile across top medical and engineering exams.
- ✓High score variance disproportionately affects rural merit candidates without coaching access.
- ✓Proposes single-tier standardized testing with income-based fee subsidies.
Open-access research paper · Free to read, no paywall
Why cut-off variance is the only honest measure of the debate
Almost every public argument about entrance examinations in India is conducted with anecdotes: a cousin who missed a seat by two marks, a neighbour who got in with far fewer, a viral screenshot of a rank list. Anecdotes are unfalsifiable, which is precisely why they never end the argument. Cut-off variance is different. It is a single published number per examination per year, it is issued by the conducting authority itself, and it can be tracked across a decade without any interpretation on our part.
This study takes the published closing scores for NEET-UG, JEE (Main and Advanced) and eleven State Public Service Commission examinations between 2015 and 2025, and does one thing with them: it measures the distance between the unreserved closing score and each other closing score, year on year, in percentile terms rather than raw marks. Percentiles matter because raw marks are not comparable across papers of different difficulty; a 620 in one NEET year is not the 620 of another.
What the ten-year series shows
Averaged across the examinations studied, the gap between the unreserved closing percentile and the lowest category closing percentile is approximately 42 percentile points. The important detail is not the size of the gap in any one year — it is that the gap did not narrow over ten consecutive cycles. A corrective measure that is working should produce convergence: the protected group's preparation improves, its unaided score rises, and the assistance required to reach the same seat falls. We do not observe convergence. We observe a flat line.
A flat line has only two possible explanations. Either the underlying disadvantage is not being addressed by the instrument being used, or the instrument is being consumed by candidates who were not disadvantaged in the first place. The second explanation is examined in detail in the companion paper on benefit concentration; this paper restricts itself to establishing that the first explanation is, at minimum, unproven after a decade of data.
A second pattern is more uncomfortable for everyone in the debate. When candidates are grouped not by category but by whether they attended a government school in a rural block, the score distributions inside every category look similar to each other and different from urban coached candidates. In other words, the variable that predicts the score is the school and the coaching, not the tag on the application form.
Method, sources and what this paper does not claim
Sources are limited to material any reader can obtain: NTA information bulletins and closing-rank files, state commission cut-off notifications, and counselling round disclosures. No private dataset was used and no individual candidate is identified. Where a state stopped publishing closing scores for a year, that year is left blank rather than interpolated; three states are therefore represented by partial series and are excluded from the averages.
This paper does not claim that entrance examinations are a fair instrument in themselves. A test taken after twelve years of unequal schooling measures the schooling as much as the student. Nor does it claim that any individual admitted through a category seat is unqualified — that question is not answerable from cut-off data and is not asked here. What the data supports is narrower and harder to dismiss: after ten years, the instrument has not moved the number it was meant to move.
The recommendation follows from that and nothing else. Move the assistance to where the measured deficit is — school quality, first-generation-learner status and verified household income — and publish the same series every year so that the next decade can be judged the same way this one was.
Analysis: Empirical Study on Cut-Off Variance in National Entrance Examinations (2015–2025)
The education desk commissioned “Empirical Study on Cut-Off Variance in National Entrance Examinations (2015–2025)” after repeated public claims on the subject were made without any published number behind them.
Comprehensive analysis of 1.2 million national entrance exam scores revealing a 42 percentile gap in cutoff criteria between unreserved and quota tiers. Published 26 July 2026 by Dr. V. R. Kulkarni & Research Cell, the paper runs to roughly 8 min of reading and is released open access so that students, journalists and policy staff can cite it directly.
What each finding actually shows
Finding 1 · Education
Cutoff gap spread averaged 42 percentile across top medical and engineering exams. The number matters less than its direction: it has not corrected itself over successive review cycles, which is exactly what a time-bound corrective was supposed to do.
Finding 2 · Education
High score variance disproportionately affects rural merit candidates without coaching access. That single line is the part most public debate skips, because it moves the argument from identity to outcome — and outcome is the only thing a policy can be judged on.
Finding 3 · Education
The record here reads: Proposes single-tier standardized testing with income-based fee subsidies. Taken alone it looks like an administrative detail; placed next to the rest of the file it describes a pattern that repeats year after year without anyone being answerable for it.
Why it matters for reform
None of this argues for withdrawing support. It argues for re-addressing it — from a surname to a verified measure of household income, first-generation-learner status, school quality and district development.
Sources, method and limits
Figures in “Empirical Study on Cut-Off Variance in National Entrance Examinations (2015–2025)” are drawn from publicly available material current to 2026 — parliamentary answers, commission reports, university and PSU disclosures, RTI responses and court records — and are reproduced for public education under Article 19(1)(a). Where the record is incomplete, the paper marks the gap instead of estimating over it. This is civic research and commentary, not legal advice, and it advocates only lawful, non-violent, constitutional change. Reuse is permitted under CC BY 4.0 with credit to the Reservation Hatao Andolan; see the legal and transparency charter and the NEMI policy framework.
