Methodology

The numbers, including the bad ones.

Most study products tell you what they got right. This page tells you what we measured, what we could not verify, and where we are most likely to be wrong. If a number here ever disagrees with something on our marketing pages, this page is correct and the other one is a bug, please report it.

What is in the corpus

5,704
Questions processed
1988–2026
Years covered
32
NCERT chapters
2,907
Compressed facts

These are NEET, AIPMT and predecessor medical-entrance papers. NEET itself only began in 2013. Anything earlier is AIPMT or its predecessors, and every question records which exam and year it came from. We do not describe this as “25 years of NEET”, because it is neither 25 years nor all NEET.

The three labels

Not every NEET question comes from one quotable NCERT sentence. Pretending otherwise would be the easiest way to make this product sound better than it is. Every question carries one of these labels, and you can see it on the question:

Verbatim line

1,579 · 27.7%

The answer appears as a continuous sentence in NCERT. We show you that sentence and highlight it in your own copy of the book.

Concept-grounded

2,908 · 51.0%

The answer is taught in an identifiable NCERT section but is not a single quotable line. It needs combining facts, reading a table, or applying a concept. We show you the section and the specific facts used.

No verified NCERT match

1,182 · 20.7%

We could not verify the answer against NCERT text. This is a real and unavoidable share of any NEET paper. We label these rather than force a match, and we would rather tell you than guess.

4,522 questions (79.3%) resolved to a specific chapter. A further 35 were matched only after our verifier disagreed with the answer key published by the source we scraped; those are flagged internally and are first in line for expert review.

How concentrated the exam really is

Measured across 4,522 mapped questions and 393 source sections:

  • Top 5% of sections carry 20.4% of questions
  • Top 10% carry 33.9%
  • Top 15% carry 44.2%
  • Top 20% carry 53.2%, the figure we quote
  • Top 50% carry 84.2%

You may have seen a claim that “15% of NCERT carries 90% of the exam”. Our own data does not support it. Reaching 90% takes roughly 60% of the book. We used to say it. We stopped.

What the audit found

A 30-question sample was pulled across chapters and difficulty bands and re-checked by hand against the printed NCERT: 24 of 30 verified (16 verbatim, 8 concept-grounded), and 6 had no verifiable NCERT source. Zero verbatim quotes were fabricated, every one we claimed was found in the book.

That sample is the part we can show you question-by-question. Beyond it, a broader expert review of the mappings has been carried out by Dr. Ambika, a PhD biologist who has coached NEET aspirants for six years, covering all three labels across the chapter set. Mappings she rejected were corrected or dropped rather than shipped.

We are still working out how much of that review to publish question-by-question. Our intent is to put the per-label error rates on this page in the same detail as the sample above, because a mapping tool that never tells you its error rate is asking for trust it has not earned.

What we exclude, and what we do not claim

  • Duplicate questions are collapsed; repeats across years keep the earliest appearance and record the rest.
  • Some diagram-dependent questions are held back from student surfaces where we cannot show the figure lawfully.
  • We do not host or redistribute NCERT textbooks. You read from your own copy; we map questions onto it.
  • We do not forecast your NEET score. We removed that feature in August 2026 because the number behind it was not validated against any real cohort.
  • We make no guarantee of any examination result.

Found a bad mapping?

Tell us and we will fix it. Email hello@neet.bio with the question and what looks wrong. Corrections are versioned, and the log will be published on this page.

Method version 1.1 · figures recomputed 5 August 2026.