The GMAT Focus Data Insights section contains a question family built entirely around a small spreadsheet. The item type is officially called Table Analysis, and it presents a sortable table — typically with five to seven columns and roughly ten to fifteen rows of data — followed by a two-part prompt. The candidate must determine which answer is true and which is false. Every other reasoning move the question demands is downstream of how that initial table is read. Most candidates who plateau in the high 500s or low 600s on Data Insights are not failing at the logic; they are reading the spreadsheet in a way that hides the relationship they need. This piece walks through the read patterns, sorting habits, and triage moves that turn Table Analysis from a guessing exercise into a controlled 90-second loop.
What a Table Analysis item actually looks like on the GMAT Focus
Each Table Analysis item gives the candidate an interactive table inside the test interface. The columns carry a mix of categorical labels (region, segment, product line) and numeric measures (revenue, cost, units sold, percentage change). The rows are individual records — a customer, a product, a quarter, a branch. The candidate can sort the table on any column, in either direction, with a single click. They can also resize columns and scroll horizontally when the table is wider than the viewport.
The prompt sits below the table and has a fixed shape. It names two statement families. Statement 1 is a claim about a subset of the data — for example, 'Among products in the Premium tier, revenue per unit increased in at least three of the last four quarters.' Statement 2 is a parallel claim that may target a different subset or a different relationship. The candidate chooses one of three options: Statement 1 is true and Statement 2 is false, Statement 1 is false and Statement 2 is true, or both statements are true.
Three structural features of the item type matter more than they look. First, the answer choices never include 'both false' — one of the two statements is always true, which collapses part of the search space. Second, the table is wide enough that not all columns fit on screen at once, so the candidate must decide which columns to anchor. Third, the prompt's two statements are deliberately designed so that a careless read of one will not contaminate the read of the other; the test is built to reward independent verification of each statement. Candidates who treat the item as 'one big question' tend to lock in early and miss the second check.
For most candidates, the table itself is roughly 8 to 15 rows. That is small. The Data Insights section gives around 2 minutes 15 seconds per item on average, and Table Analysis items can usually be closed inside 90 seconds by candidates who read the table with intent. The remaining time is better spent on the heavier items in the section — Multi-Source Reasoning and Graphics Interpretation — which tend to absorb the slack on untimed practice.
Why the prompt shape changes the read strategy
Because the two statements are independent, the candidate is never asked to compare them. The work is to verify each statement against the spreadsheet. That makes the prompt a checklist, not a debate. Treat it that way. Read Statement 1, find the rows it concerns, confirm or reject, move to Statement 2, repeat. Trying to read both at once invites an error called answer-blending, where a piece of Statement 1's logic quietly leaks into the evaluation of Statement 2.
Sorting before reading: the first 18 seconds of a Table Analysis item
Sorting is the single most leveraged move on the item type, and most candidates under-use it. The default table ordering is essentially random — set by the test designer to prevent pattern-matching on row position. The first click should almost always be a sort on whatever column the prompt's two statements share. If both statements mention 'revenue', sort by revenue. If one mentions revenue and the other mentions cost, sort on the column that appears in both. The goal is to put the relevant rows adjacent.
A useful rule of thumb: if the prompt names a single subset (for example, 'the three stores in the South region' or 'products launched before Q2'), sort the table so that subset is contiguous. A contiguous subset can be evaluated as a block. A scattered subset has to be located row by row, which adds 10 to 20 seconds per statement and breaks concentration.
Sorting direction matters as well. When a statement uses language like 'increased', 'at least', 'top three', or 'exceeded', sort in descending order. When the statement uses 'decreased', 'bottom', 'fell below', or 'no more than', sort ascending. Most candidates sort descending by default because the eye is trained to scan from the top, but a statement that asks about the smallest values can be checked faster against an ascending list — the answer is at row 1, not row 15.
Common pitfalls and how to avoid them
- Sorting on the wrong column first. If Statement 1 concerns a subset defined by region and Statement 2 concerns a subset defined by product tier, sorting on revenue gives you neither subset as a contiguous block. Read both statements before any click. Most candidates lose 20 seconds to a re-sort they could have avoided.
- Re-sorting mid-evaluation. A click on the column header is cheap, but each sort costs attention. A clean read means deciding the sort axis once and trusting it for both statements.
- Sorting by a derived value that the table does not show. The prompt may ask about a ratio or a percentage change that no column displays directly. Do not sort by the column that 'feels' related; sort by the column the statement explicitly names.
Wide tables versus tall tables: which format slows candidates down more
Table Analysis items on the GMAT Focus come in two rough shapes, and they reward different read strategies. A wide table has many columns (six or seven) and a moderate number of rows (eight to ten). A tall table has fewer columns (four or five) but more rows (twelve to fifteen). The shapes stress different parts of the read.
Wide tables are dominated by horizontal scrolling. The candidate cannot see all columns at once, and the prompt's statements may each refer to a different column. The skill is to choose a 'primary' column — usually the one named in Statement 1 — and anchor visually on it. Statement 2 may require a second anchor, and that is the moment to scroll. Many candidates try to keep both columns visible at once and lose the row-level signal. Pick the primary anchor, scroll only when forced, and return to it.
Tall tables are dominated by row count. With 15 rows, scanning for a subset takes longer, and the cost of misreading a single row compounds. The skill is to commit to the subset early. If the statement asks about 'the bottom three rows by cost', sort ascending by cost, read the top three, and write a mental note. If it asks about 'rows where Region = West and Year = Q3', do a two-pass filter: sort by Region to group the West rows, then scan those rows for Year = Q3. Two passes cost less than one tangled pass.
| Feature | Wide table | Tall table |
|---|---|---|
| Typical columns | 6–7 | 4–5 |
| Typical rows | 8–10 | 12–15 |
| Main cost | Horizontal scrolling | Row-by-row scanning |
| Best first sort | Column shared by both statements | Subset-defining column |
| Common error | Losing the row while scrolling | Misreading row 7 of 15 |
| Recovery move | Re-anchor on the primary column | Re-sort and read top to bottom |
How to verify a statement in three moves, not five
The verification step is where most candidates waste time. They read the statement, glance at the table, glance back at the statement, glance at the table again, and only then commit. That is five moves. A tighter loop uses three.
Move 1 — locate. Read the statement. Identify the subset of rows it concerns. Identify the column it asks about. Sort the table so the subset is contiguous and the column is the active sort axis. This is the only move that involves a click.
Move 2 — read the subset. Scan only the rows in the subset, top to bottom, and extract the values from the active column. Do not look at the other columns. Do not look at rows outside the subset. The narrower the visual field, the faster the read.
Move 3 — evaluate. Apply the statement's claim to the extracted values. 'At least three of the last four quarters' becomes a count: do three or more of the four values satisfy the condition? 'Revenue per unit increased' becomes a check on direction: is the second value higher than the first? Write a one-word verdict — true or false — and move to Statement 2.
That loop takes 30 to 45 seconds per statement on a well-sorted table, which leaves a comfortable 90 seconds for the full item including the initial sort. Candidates who compress the loop to two moves by skipping the locate step tend to read the wrong subset; the savings are not worth the failure rate.
Distractors that look like evidence: spotting fake patterns in the table
GMAC builds Table Analysis items around patterns that the test designer is allowed to break. A statement may say 'all four stores in the West region saw revenue grow'. The test designer will set this up so that three of the four stores grew and one was flat. A candidate reading quickly will see 'grew, grew, grew' and stop, missing the flat value at the end of the subset. The pattern is almost right; the break is the answer.
Another common distractor is the column that looks relevant but is not. A table may include a 'revenue' column and a 'revenue growth %' column. A statement that says 'revenue increased' must be checked against the absolute revenue column, not the growth column. Candidates who sort by the wrong column and find a 'yes' answer to the wrong question will mark the statement true and miss the item.
A third distractor is the row that almost fits. A subset of six rows may contain five rows that satisfy a condition and one row that is close but fails. The failure mode is rounding: the failing row is within 0.1 of the threshold, and the candidate rounds up. Read thresholds exactly. The test designer chose them precisely because they look roundable.
Common pitfalls and how to avoid them
- Stopping at the first row that confirms the claim. A statement that says 'at least three' requires all rows in the subset to be checked. Stopping at the first match is a habit from reading comprehension, where one textual anchor is usually enough. Tables do not work that way.
- Conflating two adjacent columns. A column showing 'revenue' and a column showing 'revenue per unit' sit next to each other and use the same units. A slip of the eye between them is one of the most common Data Insights errors at the 600 level.
- Forgetting that one statement must be false. Because the answer choices exclude 'both false', a candidate who finds both statements true should re-read both, not pick the first option. The 'both true' choice is real, but it is the least common of the three answers and often indicates a verification error.
Integrating Table Analysis into a Data Insights preparation plan
Table Analysis sits between Graphics Interpretation and Multi-Source Reasoning in cognitive load. Graphics Interpretation is a single chart with a fill-in-the-blank prompt. Multi-Source Reasoning pulls from three to four tabs of material. Table Analysis is the middle weight: a single dense spreadsheet and a binary-true binary-false prompt. In a preparation plan, this is the item type to drill for pace.
For most candidates, the right cadence is one Table Analysis drill per study day for the first two weeks of Data Insights prep, then every other day. The drill should use 10 to 15 items taken under timed conditions, with a 90-second target per item. After each drill, the only metric that matters is the read pattern: did the candidate sort first, locate the subset, scan only that subset, and evaluate in one pass? Items that ran over 120 seconds almost always broke one of those four moves.
The error log for Table Analysis should be different from the error log for the rest of Data Insights. Candidates who miss Multi-Source Reasoning items usually missed a piece of evidence. Candidates who miss Table Analysis items usually read the table incorrectly. The log entry should record the move that failed: 'sorted on wrong column', 'scanned outside subset', 'rounded a threshold', 'blended two statements'. Over a two-week cycle, the move that fails most often is the one to drill next.
How scoring on Table Analysis interacts with the wider Data Insights section
The Data Insights section is scored as a single number from 60 to 90 in one-point increments, derived from a section-level adaptive algorithm. Table Analysis is one of five item families in the section, and the section contains roughly 20 items total. That means Table Analysis items account for a small share of the section, but a candidate who is reliable on Table Analysis frees up attention for the heavier items.
In my experience tutoring candidates, the lift from a controlled Table Analysis loop is not just the four to five points the item family contributes directly. It is the 30 to 60 seconds per item the candidate stops losing, which gets spent on the Multi-Source Reasoning prompts where time pressure is the dominant constraint. A candidate who clears Table Analysis inside 90 seconds routinely enters the Multi-Source Reasoning block with five to eight minutes of cumulative slack, and that slack is what separates a 645 from a 685 on Data Insights.
The section's adaptive structure also rewards consistency. The first ten items of Data Insights set the difficulty band for the next ten. A candidate who is reliable on Table Analysis items contributes a steady stream of correct answers at the appropriate difficulty, which calibrates the algorithm in their favour. Candidates who gamble on Table Analysis items by skipping the sort or by blending statements produce noisy signals that push the algorithm toward easier items and cap the section score.
Building the habit: a one-week Table Analysis routine
A focused week on Table Analysis is enough to change a candidate's read pattern permanently. Day 1 is untimed: take 10 items and read the table twice — once to identify the subset, once to verify the statement. Time is not the metric. The metric is whether the second read confirms the first. Day 2 is timed: 10 items, 90 seconds each, with the stopwatch visible. Misses are acceptable; running over is not. Day 3 is an error-log day: take 5 items slowly and write a one-line note for every move that cost time. Day 4 is a 15-item timed block, 90 seconds each, and the metric is the move that failed most often on Day 3.
Day 5 is wide-versus-tall: 5 wide items, 5 tall items, each at 90 seconds. Note which shape costs more time. Most candidates lose more time on wide tables because of the horizontal scroll, and the cure is a stricter primary-anchor rule. Day 6 is integration: a 20-item Data Insights mixed block, 2 minutes 15 seconds per item, with Table Analysis items marked in the log. Day 7 is rest. The week should leave the candidate with a verified read pattern, an error log, and a measurable time-per-item drop of 15 to 25 seconds.
The single most important thing the week teaches is that Table Analysis is a control item, not a comprehension item. Candidates who treat it as 'reading comprehension on a spreadsheet' end up reading the table as a paragraph and missing the rows. Candidates who treat it as a sorting-and-verification loop end up reading the table as a database and finding the answer. The same data, the same prompt, two different scores.
Conclusion and next steps
Table Analysis is the highest-leverage item type in the GMAT Focus Data Insights section for a candidate who is plateauing in the high 500s. The read pattern is short — sort, locate, scan, evaluate — but it has to be executed in that order every time. The errors are mechanical, not conceptual, and they are the errors that respond fastest to a one-week drill routine. For candidates building a sharper preparation plan, the next step is a focused Table Analysis diagnostic: 15 timed items, an error log, and a one-week read-pattern reset. TestPrep İstanbul's diagnostic assessment is a natural starting point for candidates building a sharper preparation plan around the Table Analysis item family.