A pie chart shows City A's spending split and another shows City B's. A's food slice is 30% and B's is 25%. Which conclusion is safe?

A devotes a larger share of its spending to food than B

A's total spending is larger than B's

A's food spending is 20% more than B's

A line graph shows a country's annual GDP growth rate falling from 8% to 6% to 5% over three years. Over those years the GDP itself:

Rose each year, but by a smaller amount each time relative to its base

Cannot be determined from a growth rate

Question: what is the value of x? Statement I: x is a prime number less than 10. Statement II: x is even. Which is sufficient?

Both together, but neither alone

With 22 DILR questions in 40 minutes across five sets, which plan gives the best expected score?

Read all five, then solve the three easiest fully

Attempt all five sets, 8 minutes each

Read all five, then solve the three easiest fully

Start with set 1 and work through in order

Solve the longest set first, since it has the most questions

A table gives sales of 4,827 units against a total of 26,140. The options are 12%, 18%, 24% and 31%. The fastest correct route is:

Note that a fifth of 26,140 is about 5,228, so the answer is just under 20% — that is 18%

Long-divide 4827 by 26140 to four decimal places

Note that a fifth of 26,140 is about 5,228, so the answer is just under 20% — that is 18%

Round both to 5000 and 26000 and compute exactly

Skip the question — the numbers are too awkward

You have spent eight minutes on a set and have filled about half the grid. Seven minutes remain in the section and two unread sets are left. The right move is:

Answer whatever this set already allows, then read the two remaining sets for an easy question

Finish this set — eight minutes are already invested

Answer whatever this set already allows, then read the two remaining sets for an easy question

Abandon everything and guess on all remaining questions

Start the next set from the beginning

Reading a DILR Set — Triage, Tables and Charts

DILR gives you 22 questions in 40 minutes, delivered as four or five sets of four to six questions each. That structure is the whole strategy: you are not answering 22 independent questions, you are choosing three or four sets to commit to. A set you understand pays four marks for one piece of thinking; a set you misjudge takes fifteen minutes and pays nothing, and there is no way back into the section once the timer moves on. Most candidates who underperform in DILR do not lack technique — they picked the wrong sets in the first three minutes. This chapter is about that decision, and then about the reading skill the data-heavy sets demand: reading a table, a chart and a caselet for what the question actually needs rather than computing everything on the page.

The First Three Minutes — Choosing Sets

Spend the first two to three minutes reading every set without solving anything, and rank them. This feels like wasted time and is the highest-return three minutes in the paper: it is the difference between attempting three sets you can finish and starting one you cannot.

What makes a set worth attempting
Attempt early	Leave for last, or leave
The conditions are few and each one fixes something definite	Many conditions, each merely narrowing possibilities
You can see the shape of the answer table before starting	You cannot tell what the rows and columns should be
The numbers are small and round	Percentages of percentages, or four-figure raw data
The first question is answerable from one or two conditions	Every question needs the grid to be completely solved first
Six questions on one dataset	Four questions after a very long setup

Set a hard budget: ten minutes per set. If a set is not yielding after ten minutes, leave it. This is the single hardest discipline in the section, because a set you have invested eight minutes in feels close to done — the eight minutes are gone either way, and they are not an argument for spending seven more.
Answer the easy question inside a hard set. Many sets have one question that needs only the opening conditions. Take it, then decide whether the rest is worth the grid. Four marks for ninety seconds is a good trade even in a set you abandon.
Three sets fully solved beats five half-solved. 22 questions across five sets means a set is worth four or five marks. Completing three sets is 12–15 questions with high accuracy; touching all five is 22 questions with the accuracy of a rushed guess, and the negative marking makes that arithmetic worse than it looks.

🎯 Why triage is a skill and not a temperament

It is trainable, and the training is specific: when you practise, spend the first three minutes ranking the sets, write your ranking down, and after finishing compare it with which sets actually turned out easy. Do this ten times and your ranking becomes reliable. Practising sets without practising the choice trains half the section.

Tables — Read the Question First

A table set gives you far more data than the questions use. Computing all of it is the mistake: read the first question, find the two or three cells it needs, and touch nothing else. Then read the second question.

Read the units and the totals row before anything else. "Figures in lakh", "excluding taxes", "as a percentage of the previous year" — each of these changes every answer, and each is printed in small type above or below the table. Thirty seconds here prevents a whole set going wrong.
Percentage change is always relative to the earlier value. A rise from 40 to 50 is +25%; the fall from 50 back to 40 is −20%, not −25%. Papers print both numbers as options, so decide which value is the base before dividing.
Approximate deliberately. If the options are 12%, 18%, 24% and 31%, then 4,827 out of 26,140 does not need long division — it is a little under a fifth, so 18%. Deciding how much precision the options demand is part of reading the question.
A ratio of two ratios needs no absolute numbers. "By what percentage is A's share of the total larger than B's?" is answered from the two shares alone; the total cancels. Questions are often built so that the intimidating absolute figures are never needed.

Charts — What Each Kind Hides

Reading each chart type for its trap
Chart	What to check before using it
Bar chart	Whether the vertical axis starts at zero. A truncated axis makes a 3% difference look like a doubling.
Stacked bar	Whether each bar totals 100% (shares) or a real quantity (amounts). The two need different arithmetic.
Pie chart	A pie shows shares only. Two pies of different totals cannot be compared slice to slice without the totals.
Line graph	Whether the line is a quantity or a growth rate. A falling growth rate still means the quantity is rising.
Two axes on one chart	Which series belongs to which axis. This is the most reliable trap in the whole topic.

⚠️ A share can rise while the amount falls

If a company's share of a market goes from 20% to 25% while the market shrinks from 1000 to 700, its sales fell from 200 to 175. Questions are built on exactly this gap between a percentage and a quantity, so always ask which of the two the question is about.

Caselets and Data Sufficiency

A caselet is a table in prose. The data arrives as sentences, and the first move is always the same: build the table the sentences describe before answering anything. Trying to answer from the paragraph is how candidates lose caselets they could have solved.
Mark what is fixed and what is a range. "At least 20 students" is not "20 students". Write ≥ 20 in the cell. Half the wrong answers in caselets come from treating a bound as an equality.
Data sufficiency asks whether you can answer, not what the answer is. Test each statement alone, then both together, and stop as soon as the question is decided. Computing the actual value wastes the time the format was designed to save you.
"Unique answer" is the test, not "some answer". A statement that narrows the possibilities to two is not sufficient. This is the single most common data-sufficiency error, and it comes from reading "I have learnt something" as "I can answer".

Key takeaways

Read every set before solving any of them, and rank them. Three minutes of triage is worth more than three minutes of solving.
Ten minutes per set is the budget. Time already spent is not a reason to spend more — that money is gone either way.
Three sets solved fully beats five touched. Negative marking makes a rushed guess cost more than the mark it might have won.
Read the units, the totals row and the footnote before the data. "Figures in lakh" changes every answer in the set.
A percentage change is relative to the earlier value: 40 → 50 is +25% but 50 → 40 is −20%. Both appear as options.
A rising share does not mean rising sales — the total may have shrunk. Always ask whether the question is about a percentage or a quantity.
In data sufficiency, a statement that leaves two possibilities is not sufficient. The test is a unique answer, not progress towards one.

Practice questions (8)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

A quantity rises from 40 to 50 in one year and falls back to 40 the next. The two percentage changes are:
1. +25% and −25%
2. +25% and −20%
3. +20% and −20%
4. +20% and −25%
Show answer
Answer: B — +25% and −20%
A change is always divided by the earlier value. Up: 10/40 = +25%. Down: 10/50 = −20%. The asymmetry is the point — a rise and the fall that undoes it are never the same percentage, and a paper that offers "+25% and −25%" is testing exactly that.
A company's market share rises from 20% to 25% while the market shrinks from ₹1000 crore to ₹700 crore. Its sales:
1. Rose by 25%
2. Fell by 12.5%
3. Stayed the same
4. Rose by 5%
Show answer
Answer: B — Fell by 12.5%
Sales are share × market: 20% of 1000 = ₹200 crore, then 25% of 700 = ₹175 crore. The fall is 25/200 = 12.5%. So the share rose and the sales fell — the gap between a percentage and a quantity, which is the most reliable trap in data interpretation. Option A reads the share change as the sales change.
A pie chart shows City A's spending split and another shows City B's. A's food slice is 30% and B's is 25%. Which conclusion is safe?
1. A spends more on food than B
2. A devotes a larger share of its spending to food than B
3. A's total spending is larger than B's
4. A's food spending is 20% more than B's
Show answer
Answer: B — A devotes a larger share of its spending to food than B
A pie chart shows shares and nothing else. Without the two totals, no comparison of absolute amounts is possible — B could be a far larger city, in which case 25% of its spending exceeds 30% of A's. Only the statement about shares is safe. Options A, C and D all quietly assume the totals are known or equal.
A line graph shows a country's annual GDP growth rate falling from 8% to 6% to 5% over three years. Over those years the GDP itself:
1. Fell each year
2. Rose each year, but by a smaller amount each time relative to its base
3. Stayed roughly flat
4. Cannot be determined from a growth rate
Show answer
Answer: B — Rose each year, but by a smaller amount each time relative to its base
A positive growth rate means the quantity is still growing; a falling growth rate only means it is growing more slowly. All three rates are positive, so GDP rose every year — by 8%, then 6%, then 5% of the preceding year. Option A confuses a falling rate with a falling quantity, which is the standard misreading of this chart type. Note also that the absolute increase need not shrink: 5% of a larger base can exceed 6% of a smaller one.
Question: what is the value of x? Statement I: x is a prime number less than 10. Statement II: x is even. Which is sufficient?
1. Statement I alone
2. Statement II alone
3. Both together, but neither alone
4. Neither, even together
Show answer
Answer: C — Both together, but neither alone
Statement I allows 2, 3, 5 or 7 — four possibilities, so not sufficient. Statement II allows every even number, so certainly not sufficient. Together: the only even prime is 2, so x = 2 uniquely, and both together are sufficient. The trap is answering A because statement I "nearly" determines x — narrowing to four values is not answering.
With 22 DILR questions in 40 minutes across five sets, which plan gives the best expected score?
1. Attempt all five sets, 8 minutes each
2. Read all five, then solve the three easiest fully
3. Start with set 1 and work through in order
4. Solve the longest set first, since it has the most questions
Show answer
Answer: B — Read all five, then solve the three easiest fully
Reading all five costs about three minutes and tells you which are tractable; three sets solved fully is 12–15 questions at high accuracy, which beats 22 rushed ones once negative marking is counted. Option A spreads the time so thin that no set gets finished. Option C ignores that sets are not ordered by difficulty. Option D confuses length with value — a long setup with four questions is worse value than a short one with six.
A table gives sales of 4,827 units against a total of 26,140. The options are 12%, 18%, 24% and 31%. The fastest correct route is:
1. Long-divide 4827 by 26140 to four decimal places
2. Note that a fifth of 26,140 is about 5,228, so the answer is just under 20% — that is 18%
3. Round both to 5000 and 26000 and compute exactly
4. Skip the question — the numbers are too awkward
Show answer
Answer: B — Note that a fifth of 26,140 is about 5,228, so the answer is just under 20% — that is 18%
The options are 6 percentage points apart, so an estimate to within 2 points settles it. A fifth of 26,140 is 5,228; our figure of 4,827 is smaller than that, so the share is below 20%, and 18% is the only candidate. (The exact value is 18.47%.) Option A wastes forty seconds producing precision the options do not use — and reading the options before calculating is the habit this question is really about.
You have spent eight minutes on a set and have filled about half the grid. Seven minutes remain in the section and two unread sets are left. The right move is:
1. Finish this set — eight minutes are already invested
2. Answer whatever this set already allows, then read the two remaining sets for an easy question
3. Abandon everything and guess on all remaining questions
4. Start the next set from the beginning
Show answer
Answer: B — Answer whatever this set already allows, then read the two remaining sets for an easy question
A half-filled grid usually answers at least one question already — take those marks first, since they are nearly free. Then look at the unread sets, because an easy question in an unseen set is worth more than the seventh minute of a stubborn one. Option A is the sunk-cost error: the eight minutes are spent whatever you do next and are not a reason to spend seven more. Option C throws away marks the grid has already earned, and negative marking makes blind guessing worse than not answering.