6,580 facts: what our NEET data is showing us

PrepArena research notes · 10 September 2026 · Prepared with AI assistance

Can we break NEET preparation into small learning goals, so a student knows exactly what to work on next?

We looked inside PrepArena's own content to explore that question. We counted 6,580 fact entries across 79 chapter fact lists, and checked how often those entries had been linked to past-year questions (PYQs). Some results were encouraging. Others challenged our first ideas.

Here is what we found, what the numbers mean, and how they are shaping our plans to improve PrepArena.

What did we count?

We checked content version 2026.09.09-1411. Its question bank contains 5,917 question items: 4,978 authored items and 939 PYQs. Alongside it are the 6,580 fact entries. A “fact entry” here means a small piece of content in our system—a definition, a mechanism, a number or a step. It is not automatically a complete learning goal or an independently checked truth.

4,386 question items—about 74%—already have fact labels. That could help us connect a wrong answer to the idea a student needs to revisit. But a label alone does not prove the connection is correct, or that a student has learned the idea.

1. NEET content is more than a list of definitions

The largest group in our fact lists is definitions: 2,413 entries, or about 37%. Mechanisms—descriptions of how something happens—make up another 22%. Process steps account for about 10%.

The table shows all seven groups. A recorded match means our system linked a fact entry to a PYQ. The last column is the average number of these links per entry, including entries with no links. Neither column measures exam marks.

The “number” group has the highest average, about 0.70 matches per entry. That caught our attention: values and quantities should not be ignored when thinking about question-solving practice. It does not mean memorising numbers is the best use of every student's time.

Our original comparison grouped mechanisms and process steps together, and compared them with definitions, numbers and examples. Their averages were very close: 0.56 versus 0.57 matches per entry. That comparison left out exceptions and comparisons; adding exceptions to the second group gives about 0.56.

So these counts do not give us a sound reason to favour one group as a route to more marks. Nor do they prove that the groups are equally useful. The labels and the links need checking first. Biology, Chemistry and Physics all involve both knowledge and reasoning; one fixed study method for each subject would be too simple.

2. A short “top facts” list does not tell the whole story

We sorted the entries from most recorded matches to least. Then we asked: how many entries account for most of those matches?

The top 300 entries account for 31% of recorded matches. To reach 80%, we need 1,525 entries—about 23% of the full fact list.

This is useful for exploring what to review first. It does not justify “learn these facts and cover 80% of NEET.” One question can match several facts, and some links may be wrong. The chart shows shares of 3,689 stored links, not shares of questions or marks.

Also, only 2,262 of the 6,580 entries have any recorded match. The others may include topics the system missed. A zero is not a reason to skip a topic in the official syllabus.

3. We need to fix the gaps before drawing bigger conclusions

Rechecking the counts reproduced the headline numbers. That is useful, but counting stored links is different from checking whether each link makes sense.

• The matching needs review. It is based on text similarity. Similar words do not always mean a question tests the same idea. We need subject-reviewed examples to measure missed and incorrect links.
• 1,372 entries—about 21%—are marked “disputed.” In this system, that means the automated reviewers disagreed. It does not mean 21% are wrong. It does mean these entries need a clear review decision.
• The year ranges do not line up. The published PYQ bank contains questions from 2016–2023. The fact lists say 2016–2026. We need to trace the actual matching inputs before claiming which papers the match counts cover. The counts cannot be presented as an audit of every recent exam.

Past papers also cannot replace the official syllabus (2026 reference). Students should use the syllabus for their own exam year.

How this is shaping PrepArena

These findings point us toward three priorities:

1. Make the content links more trustworthy. Check the fact-to-question matches and resolve content disagreements with qualified subject review.
2. Make the next learning step clearer. Explore smaller goals that include remembering, explaining and applying an idea. The current question metadata records no prerequisite links; those relationships need careful work rather than automatic assumptions.
3. Test learning, not just activity. Check whether students can solve new NEET-style questions and remember the ideas later. A completed set or a correct answer just after reading a solution is not enough.

PrepArena already has answer and review history, with basic fact progress derived from it. What we still need is evidence that our progress measures reflect lasting learning. The improvements above are research and product priorities, not a claim that a complete personalised course is ready.

Our vision remains free, high-quality NEET preparation for every student, including students in smaller towns working toward an affordable government MBBS seat. Sharing what the data supports—and where it challenges us—is part of building toward that vision.

You can explore our current free study guides and practice. This article is available in Hindi too; linked resources may be in English.