The CSAT trap: how the qualifying paper eliminates more aspirants than it should

Every year, aspirants who would have cleared GS Paper 1 comfortably are eliminated because they failed CSAT. This is not a myth or a coaching institute scare tactic — it is a documented pattern. The qualifying threshold is 33%, which sounds low. What sounds low and what is achievable under exam pressure on a paper that is getting harder are two different things.

This article uses four complete years of tagged CSAT data — 2022, 2023, 2024, and 2025 — to show exactly what the paper looks like, which section is driving the difficulty increase, and what aspirants consistently underestimate.

Four complete years of CSAT difficulty data

With all four years now tagged, the trend is unambiguous. Easy questions have declined in every year. Hard questions have grown in every year. The paper is not "getting a little harder" — it is systematically increasing in difficulty at a consistent rate.

Which section is driving this

The data is unambiguous on which section is responsible: Numerical Ability has by far the highest Hard question proportion of any CSAT section. The shift from arithmetic (percentages, simple profit/loss) to number theory, algebraic reasoning, and abstract mathematical patterns has made this section categorically harder than it was in 2022.

Data Interpretation & Sufficiency is the second-hardest section — not in absolute Hard question count, but in the complexity of calculation required. Recent DI questions require two or three dependent calculations rather than one.

Logical & Analytical Reasoning sits in the middle — broadly medium difficulty, but with the highest variance. Some reasoning questions are among the easiest in the paper; others are among the hardest.

Reading Comprehension remains the most accessible section. The difficulty has crept up slightly — passages are longer, abstract topics have appeared — but RC is still the section where a well-prepared aspirant can score most reliably.

The trap — what aspirants consistently underestimate

The 33% qualifying threshold creates a false sense of safety. Here is the calculation that reveals the trap:

• 80 questions × 2.5 marks each = 200 marks total
• 33% threshold = 66.67 marks = approximately 27 questions correct with zero negatives
• Each wrong answer costs 0.83 marks (one-third of 2.5)

Now consider a typical under-prepared aspirant's exam:

• They attempt 65 questions (the usual advice is "attempt as many as possible")
• They get 40 correct and 25 wrong
• Score: (40 × 2.5) - (25 × 0.83) = 100 - 20.75 = 79.25 marks — comfortably qualifying

Now consider the same aspirant on the 2025 paper, which has more Hard questions:

• They still attempt 65 questions
• They get 35 correct and 30 wrong (the Hard questions they guessed on cost them)
• Score: (35 × 2.5) - (30 × 0.83) = 87.5 - 24.9 = 62.6 marks — below the threshold

The difference is 4–5 wrong answers on Hard questions they should have left blank.

The section-by-year evolution

Numerical Ability has grown in both volume and difficulty — a double squeeze. Reading Comprehension has held broadly stable. Reasoning has slight year-to-year variation but no clear trend. DI has stayed broadly constant in volume.

The correct strategy — section-specific

Reading Comprehension

RC is the section where accuracy matters most. Every RC question you answer correctly at high accuracy is a mark you have not risked. Strategy: read each passage fully before attempting questions, answer only from the text (do not bring outside knowledge), and flag any question where you are unsure — RC wrong answers are just as expensive as Numerical wrong answers.

Logical & Analytical Reasoning

This section rewards pattern recognition developed through practice. Strategy: do timed practice sets, not untimed study. Under exam conditions, reasoning questions that take more than 90 seconds should be left for a second pass or skipped. The time cost of a Hard reasoning question is often not worth the mark.

Numerical Ability

This is the section where the trap is set. Strategy: before the exam, know your personal difficulty threshold. Practise enough that you can reliably identify, within 30 seconds of reading a question, whether it is within your solving ability. Questions above that threshold should be skipped — not attempted with a guess. One confidently wrong Hard Numerical question costs the equivalent of a correct Easy question.

Data Interpretation

DI questions have a time cost that is easy to underestimate. Strategy: attempt DI questions on a second pass after completing RC and Reasoning. Calculate the number of marks you need from DI before attempting — if you are already above the threshold from RC and Reasoning, the risk-reward of DI Hard questions may not be worth it.