Data Interpretation — XAT Quantitative Ability & Data Interpretation

A DI set is not a comprehension test in disguise — the data is usually laid out plainly enough that misreading it is avoidable with care. The real test is arithmetic economy: can you extract exactly the numbers a question needs, skip the ones it doesn't, and estimate instead of computing exactly when the five options are far enough apart that estimation is all that's required?

1. What XAT actually asks

Data Interpretation is estimated at roughly 4-6 of QA&DI's 28 questions across recent papers (see docs/exam-briefs/xat-2026-brief.md), usually presented as one or two data sets (a table, a bar chart, a line graph, or a pie chart) with 3-5 questions attached to each. Every question carries five options and the standard +1/-0.25 marking, exactly as elsewhere in QA&DI.

DI sets test three distinct skills, and most questions lean on one specific skill more than the others: accurate extraction (reading the correct row, column, or bar without error), derived computation (combining two or more extracted values — a ratio, a percentage, a difference), and estimation under time pressure (recognising when an exact calculation is unnecessary because the options are far enough apart).

2. Reading the data set before touching any question

1. Read the title, units, and axis labels first. A value in "₹ lakh" versus "₹ crore," or a percentage versus an absolute figure, changes every subsequent calculation — this single step prevents the most common DI error category.
2. Scan for a footnote or an unusual entry. Data sets sometimes include a category marked "others" or a note about a base-year change — these frequently matter for at least one question in the set.
3. Do not try to memorise the whole table. Note only its overall shape (how many categories, how many years or periods) and return to the table for exact figures per question — trying to hold every number in memory wastes time and invites recall errors.

3. Percentage and ratio questions — the fastest sub-type

The majority of DI questions ask for a percentage change, a percentage share, or a ratio between two extracted values — exactly the arithmetic from the Arithmetic chapter, applied to table data instead of a word problem.

$\text{percentage share}=\frac{\text{part}}{\text{whole}}\times 100,\text{percentage change}=\frac{\text{new}-\text{old}}{\text{old}}\times 100$

The base-identification discipline matters even more here than in a standalone arithmetic question, because a DI table often has multiple plausible "wholes" (a row total, a column total, a grand total) — always confirm exactly which total a percentage share question intends before dividing.

4. Estimation — when exact computation is unnecessary

XAT's five options are sometimes spaced closely enough that only an exact calculation resolves the question — but often they are not, and a rough estimate reaches the answer faster.

The test: compare the gap between the closest two options to the likely error in a quick mental estimate. If the options are ₹120, ₹125, ₹340, ₹360, ₹580, a rough calculation immediately separates the answer into one of three well-separated clusters, and only a coarse estimate is needed. If the options are ₹342, ₹345, ₹348, ₹351, ₹354, exact computation is unavoidable.

Rounding for speed: round each extracted figure to 2-3 significant figures before combining them, unless the options are tightly clustered — this single habit is usually enough to answer widely-spaced-option questions correctly in a fraction of the full computation time.

Table — Annual sales (₹ crore) of four product lines at a company, 2021-2024:

6. Common traps

• Misreading the axis or table unit (lakh vs. crore, percentage vs. absolute value) — always confirm units before any calculation, not after getting a suspiciously large or small answer.
• Using the wrong "whole" for a percentage-share question when a table offers multiple plausible totals (row, column, grand total).
• Reporting a raw difference as if it were a percentage — always divide by the correct base before treating a number as "the percentage."
• Assuming the category with the highest absolute value also has the highest percentage share — as Q2 shows, these can diverge when the total itself changes across categories being compared.
• Inverting a requested ratio's direction — always match the order the question states (X to Y, not Y to X).
• Computing exactly when the options are widely spaced enough for a fast estimate, wasting time the section's tight budget doesn't allow.

7. When to guess, and why

Within your first 8 skips across all of Part 1, a DI question requiring a slow, error-prone multi-step calculation on closely-spaced options is a reasonable one to skip, particularly late in a DI set where earlier questions may have already consumed more time than budgeted. Beyond your 8th skip, a blind 1-in-5 guess has an expected value of exactly 0, while a blank costs -0.10 — mark your best remaining guess.

Because DI options are frequently derived from a shared, extractable data point, even a partial extraction (confirming just one or two figures from the table) often eliminates two or three clearly-inconsistent options before a final guess is needed.

Summary

• Data Interpretation is roughly 4-6 of QA&DI's 28 questions, usually 1-2 data sets with 3-5 questions each.
• Three distinct skills: accurate extraction, derived computation (ratios, percentages, differences), and estimation under time pressure.
• Always read titles, units and axis labels before any calculation — unit confusion (lakh vs. crore) is the most preventable, most common DI error.
• Confirm exactly which "whole" a percentage-share question intends when a table offers multiple plausible totals.
• Compare the gap between the closest options to your likely estimation error — widely-spaced options rarely need exact computation.
• The category with the highest absolute value doesn't always have the highest percentage share — these can diverge when the underlying totals differ.
• Always match a requested ratio's stated direction (X to Y is not Y to X).
• Past your 8th free skip in Part 1, even a partial data extraction often eliminates enough options to make a late guess meaningfully better than blind chance.