Caselet Data Interpretation for CAT: The 4 Sweep Pass

DILR

A bar chart with twelve bars takes you forty seconds to read and you feel in control. The same twelve numbers written as a paragraph takes four minutes and you are not sure you have them all. Nothing about the data changed. What changed is that you now have to build the chart yourself, under a clock, while reading.

That is the entire difficulty of a caselet, and it is a transcription problem rather than a reasoning one. Aspirants who lose marks on caselet style data interpretation questions for CAT almost never fail at the arithmetic. They fail because they started calculating before the paragraph had been turned into a table.

Why a Caselet Feels Harder Than a Table

A chart hands you the structure and asks you to read values out of it. A caselet hands you values and asks you to infer the structure. The second is strictly more work, and CAT knows it, which is why caselets are used to raise difficulty without raising the mathematics.

The cost shows up in a specific way. Reading prose is sequential, so you meet a number about department C before you know that there are four departments, and you have nowhere to put it. You hold it in memory, meet three more, and by sentence five you are rereading rather than reading.

• You started reading without a structure, so early numbers had nowhere to live.
• You wrote numbers in a list rather than a grid, so checking a relationship means scanning.
• You converted a percentage before knowing what it was a percentage of.
• You reread the paragraph three times, which is the real time cost rather than any calculation.

What a Caselet Is Actually Hiding

Every caselet contains the same four ingredients, differently disguised. Naming them as you read is what turns prose back into a grid.

Ingredient	What it looks like in prose	Where it goes
Entities	Four departments, five cities, three years	The rows and columns of your table
Absolutes	The company employs 400 people	Straight into a cell or a total
Relatives	B has twenty more than A, C is half of B	An expression, resolved once A is known
Constraints	The four departments account for everyone	A totals row that closes the table

Reading for one ingredient at a time is what makes the paragraph tractable. Trying to catch all four on a single read is exactly the behaviour that produces three rereads, because each ingredient wants a different kind of attention.

Constraints are the ingredient aspirants skip, and they are usually what makes the set solvable. A sentence saying the categories are exhaustive is not filler. It is the equation that lets you find the last unknown by subtraction.

The Extraction Pass: 4 Sweeps

Four passes over the same paragraph, each looking for one thing. Reading four times sounds slow and is faster, because each pass is quick and none of them requires holding anything in memory.

Why Four Passes Beat One Careful One

A single careful read asks you to do four jobs at once while the structure is still forming. Four narrow reads ask you to do one job each against a structure that improves every time. The total reading is longer and the total time is shorter, which is counterintuitive enough that most aspirants never try it.

Sweep 1: Name the Entities and Draw the Grid

Read the paragraph once for nouns. How many things are being compared, and along how many dimensions. Then draw the empty table before you look at a single number.

Most caselets are two dimensional: categories down the side, attributes across the top. If a third dimension appears, such as two years, draw two tables rather than one crowded grid. Two clean tables are read faster than one that needs interpreting.

• Count the categories. If the paragraph names four departments, draw four rows plus a totals row.
• Count the attributes. Headcount, revenue, percentage of women, whatever is measured.
• Add a totals row and a totals column even if the set never mentions them.
• Leave the grid deliberately roomy, since caselet questions often add a derived column later.

Sweep 2: Place the Absolute Numbers

Second read, looking only for plain values. Totals, counts, and any figure stated without reference to another figure. These go straight into cells, and they are the anchors everything else resolves against.

Percentages of a known total count as absolutes here, so convert them immediately. Twenty five percent of four hundred is one hundred and belongs in a cell as one hundred. Leaving it as a percentage means converting it later, usually twice.

• Plain counts and totals go straight into cells.
• Percentages of a known total are absolutes in disguise. Convert them now.
• Percentages of an unknown base wait for sweep three, since they are relatives.
• Averages are absolutes only if the count is known, so check the count before placing one.

Sweeps 3 and 4: Relatives, Then Unknowns

The last two sweeps run together, because a relative statement you cannot yet resolve becomes a marked unknown rather than a problem. Separating them is what stops you stalling on a chain that needs an anchor you have not reached.

Sweep 3: Convert the Relative Statements

Third read, looking for comparisons: more than, twice, half of, exceeds by, in the ratio of. Each becomes an expression rather than a value, and expressions resolve in a chain once one anchor is known.

Write them as expressions in the cell rather than solving each as you meet it. A chain of three relatives usually collapses in one step once the anchor arrives, and solving them individually as you read means solving the same chain three times in three different orders.

The Four Relative Phrasings That Cause Errors

• More than versus times. Twenty more than A is A plus 20. Twenty times A is very different, and the wording is often close.
• Percentage of versus percentage more than. Sixty percent of A is 0.6A. Sixty percent more than A is 1.6A.
• Ratio order. The ratio of A to B being 3:5 is not the same as B to A, and caselets state it both ways within one paragraph.
• Exceeds by. Always a difference, never a factor, however the sentence is arranged.

Sweep 4: Mark What Is Still Unknown

Fourth read is the shortest. Look at the grid, circle every empty cell, and for each one ask which constraint would close it. Usually the answer is the totals row, which is why you drew it in sweep one.

• Circle every empty cell rather than mentally noting it.
• Against each, write which row or column total would close it.
• If nothing closes it, mark it as genuinely undetermined and move on.
• Count the circles. One circle means subtraction closes the grid. Three or more means a question will ask what can be determined.

Some cells stay genuinely unknown, and knowing which ones is worth marks by itself. Caselet questions frequently ask what can be determined, and an aspirant with a marked grid answers that instantly while everyone else recalculates.

A Worked Caselet, Sweep by Sweep

A company has four departments, A, B, C and D, employing 400 people in total. Department A has 25 percent of the total. B has 20 more employees than A. C has half as many as B. Every employee belongs to exactly one department.

1. Entities. Four departments plus a total. Draw five rows and one headcount column.
2. Absolutes. Total is 400. A is 25 percent of 400, so A is 100. Two cells filled.
3. Relatives. B is A plus 20, so 120. C is half of B, so 60.
4. Unknowns. Only D is empty. The exhaustive constraint closes it: 400 minus 100 minus 120 minus 60, so D is 120.
5. Read the grid. A is 100, B is 120, C is 60, D is 120. B and D are equal, which is exactly the kind of thing a question will ask.

The grid took under a minute and now answers every question in the set without rereading the paragraph once. Compare that against solving question by question from the prose, which rereads the same four sentences four times.

Building the Extraction Pass Into Your CAT Preparation

Extraction is trainable separately from solving, and separating them is what makes it fast. Three short blocks a week make the four sweeps automatic inside a month.

1. Block one, structure only. Take five caselets and draw the empty grid for each. Fill nothing, answer nothing.
2. Block two, extract only. Take the same five and complete the grids. Still answer no questions.
3. Block three, full sets under time. Solve three complete caselets and record how long the extraction took.

Then compare against the tabular version of the same skill. A round of data table practice questions shows you how much faster you read a grid you did not have to build, which is a useful measure of what the extraction is buying. Real papers are the final check, and topic-wise CAT exam previous year questions write caselets far less tidily than practice sets do.

Keep the written reference minimal while you drill. The free CAT DILR notes and cheatsheets already carry the standard conversions, and the AI study planner for CAT 2026 will hold the three blocks in a week that keeps filling up.