GMAT Table Analysis: how to sort, calculate and judge each statement

Updated 25 September 2026 · 7 min read

Table Analysis is the most "spreadsheet-like" question on the GMAT. You get a table of data that you can sort by any column, and a set of statements to judge one by one. There is rarely any hard mathematics. Points are lost to misreading a column, to a careless median, or to treating "consistent with the data" as "proven by the data".

This guide explains the format, gives a method you can use on every table, and works through an original example with three statements, followed by the traps to watch for.

What Table Analysis looks like

Table Analysis is one of five question types in the GMAT Focus Data Insights section (20 questions, 45 minutes, on-screen calculator available). GMAC describes it as a table you sort to organise the data, so you can determine whether certain conditions are met. The statements that follow have opposing answer options, such as Yes/No or True/False.

A typical question has:

• A short introduction explaining what the table shows, sometimes with notes on units or definitions.
• A table with several columns and anywhere from a handful to a few dozen rows. Clicking a column heading sorts by that column.
• Three statements, each with two options, such as "Yes" and "No" or "Would help explain" and "Would not help explain".

Prep companies consistently report that you must judge every statement correctly to earn credit for the question, so one careless statement costs the whole question.

A five-step method

1. Read the introduction and the column headings before anything else. Note units (thousands? percent?), what one row represents, and any definitions. Twenty seconds here prevents most errors.
2. Read the statement's exact claim. Is it about a count ("more than half"), a rank ("the third highest"), a calculation ("the average"), or a proof ("must be true")?
3. Choose the column to sort by. Rank and median statements become trivial after one sort. Statements that combine two columns ("revenue per employee") need a quick calculation, which is what the calculator is for.
4. Decide, then move on. Do not reread the table for each statement from scratch; each statement is independent.
5. Re-sort for the next statement. A sort changes the whole view. Never assume the order from the previous statement still holds.

Worked example: a retail chain's stores

The table shows data for the eight stores of a retail chain in 2026: the region, annual revenue in thousands of dollars, the number of employees, and the percent change in revenue compared with 2025.

For each statement, select Yes if it is true based on the information provided, or No otherwise.

1. The median 2026 revenue of the eight stores is greater than $800 thousand.
2. The store with the highest 2026 revenue per employee is in the West region.
3. Dogwood's 2025 revenue was $400 thousand.

Statement 1: the median

Sort by revenue: 480, 560, 610, 720, 840, 930, 1,020, 1,150. With eight values, the median is the average of the 4th and 5th: (720 + 840) ÷ 2 = 780. That is less than 800, so the answer is No.

The trap is picking one "middle" value. 840 is the 5th value and is above 800; a hurried reading picks it and answers Yes. With an even number of rows, always average the two middle values.

Statement 2: revenue per employee

This combines two columns, so you calculate. Revenue per employee, in thousands of dollars:

Hazel is highest, and Hazel is in the West. The answer is Yes.

Notice the shortcut: the store with the most revenue (Cedar) is not the answer, and the question is built so that the biggest store looks right. You do not need to calculate every row, though. Stores with high revenue and few employees are the only candidates, so Hazel (7 employees) and Gum and Cedar are the ones to check.

Statement 3: reverse percentages

Dogwood's revenue rose 20 percent to 480. So 2025 revenue × 1.20 = 480, and 2025 revenue = 480 ÷ 1.2 = 400. The answer is Yes.

The trap is subtracting 20 percent of 480 (which gives 384). A percent change is always a percentage of the earlier value, so divide by the multiplier rather than subtracting.

Answers: 1 No, 2 Yes, 3 Yes.

An extra statement to try

More than half of the stores whose revenue grew in 2026 had at least 12 employees. Filter mentally for positive changes: Alder, Cedar, Dogwood, Fir and Gum (five stores). Of these, Alder (12), Cedar (16) and Gum (14) have at least 12 employees: three of five, which is more than half. Yes. Note that "at least 12" includes Alder's exactly 12, and that negative changes such as Hazel's −1 are not growth.

The traps that cost the most points

A special case deserves attention: statements that ask whether data "would help explain" something, or whether a claim "can be determined". Here you are not calculating a result; you are judging whether the table contains enough information. A table that shows totals by region, for example, cannot determine anything about an individual store within a region.

Using the calculator and sorting efficiently

• Estimate first. In Statement 2, only a few stores could possibly lead, so only a few divisions matter.
• Use the calculator for close calls. Gum (72.9) and Fir (72.0) are close; a quick calculation settles which is higher if a statement depended on it.
• Sort once per statement, and sort by the column the statement is about. For "the third highest" or "the median", sorting is faster and safer than scanning.
• Count rows deliberately. When a statement says "more than half", write down the total number of qualifying rows before you compare.

How to practise Table Analysis

• Practise with real sorting when you can. Table Analysis in the exam interface is interactive. Our Table Analysis practice questions show the full data table and explanations that say which column to sort and what to compute.
• Drill the underlying arithmetic. Percent change, averages and medians do most of the work; our percent and statistics sets build that fluency.
• Log the cause of every miss. In Table Analysis, the cause is usually a reading error, not a maths error, and an error log makes that pattern visible.

Table Analysis sits beside Two-Part Analysis, Graphics Interpretation and Multi-Source Reasoning in Data Insights; the Data Insights overview covers all five formats and a pacing plan for the section.