GMAT · Guides · October 30, 2026 · 12 min read
GMAT Data Insights: the five question types, calculator, scoring, timing and worked examples

GMAT Data Insights (DI) is a 45-minute section of 20 questions that test how you use numbers, tables, charts and text together to reach a decision. It has five question types: Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphics Interpretation and Two-Part Analysis. DI is the only section with an on-screen calculator, and many of its questions have several parts, which must all be right for the question to count. Below is how the section works, how it is scored, a timing plan and a worked example of each type, all written by us.
Data Insights at a glance
As of October 2026, per mba.com and gmac.com:
| Data Insights | |
|---|---|
| Questions | 20 |
| Time | 45 minutes (2 min 15 s per question on average) |
| Question types | Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphics Interpretation, Two-Part Analysis |
| Calculator | on-screen calculator allowed |
| Content | maths, data analysis, verbal reasoning, or all three in one question |
| Multi-part questions | all parts must be correct to score |
| Section score | 60–90, one third of the 205–805 total |
| Review & Edit | bookmark questions, then change up to 3 answers if time remains |
DI carries the same weight in the total score as Quant and Verbal, and you can take it first, second or last. mba.com does not publish a fixed number of questions for each type, so expect the mix to vary. For the whole exam structure, see the GMAT syllabus and exam pattern.
How Data Insights is scored
All or nothing on multi-part questions. GMAC’s own Data Insights guide on gmac.com states that some DI questions have multiple parts and that you need to get all of the parts correct to score. A Table Analysis question with three statements, a Graphics Interpretation question with two drop-downs and a Two-Part Analysis question with two columns are each one question: two out of three right counts the same as zero.
What decides the score. mba.com says GMAT scores depend on the difficulty and other characteristics of the questions you answer correctly and incorrectly, and on the number you leave unanswered. It also lists “certain question types are weighted more heavily than other question types” among scoring myths, so a quick Data Sufficiency question is not worth less than a long Multi-Source Reasoning set.
Unanswered questions are penalised. If you do not finish the section, your score is reduced in line with the number of questions left blank. A guess is always better than a blank.
GMAC does not publish the scoring formula, so any score calculator is an estimate. Our GMAT score calculator and score chart show how a DI section score combines with Quant and Verbal.
The on-screen calculator
The calculator is a help, not a strategy. Clicking numbers on screen is slower than it seems, and most DI questions are designed so that estimation or simple arithmetic is enough.
- Use it for ugly divisions and percentages: revenue per employee, percent change between two four-digit numbers, a weighted average across rows.
- Estimate first. If the answer choices or drop-down options are far apart, a rough calculation decides the question faster than exact keystrokes.
- Do not use it for Data Sufficiency in most cases. DS asks whether an answer can be found, not what it is.
The five question types, with worked examples
All examples are original questions written for this guide, not official GMAT questions. On the real test, tables and charts are interactive screens; here they are shown as text.
1. Data Sufficiency
A question followed by two statements. You decide whether the statements, alone or together, give enough information to answer, choosing from the same five answer choices every time: (A) statement 1 alone is sufficient, (B) statement 2 alone, (C) both together but neither alone, (D) each alone, (E) not even both together.
A bakery sold 50 items yesterday, all of them either muffins at $3 each or cookies at $2 each. How much revenue did the bakery receive from muffins yesterday?
(1) The bakery’s total revenue from muffins and cookies yesterday was $130.
(2) Yesterday the bakery sold 10 more muffins than cookies.
Statement (1). With m muffins and c cookies, m + c = 50 and 3m + 2c = 130. Substitute c = 50 − m: 3m + 100 − 2m = 130, so m = 30 and muffin revenue is $90. One value: sufficient.
Statement (2). m = c + 10 and m + c = 50, so 2c + 10 = 50, c = 20, m = 30: again $90. Sufficient.
Answer: D. The trap is C: assuming you need both statements because the question involves two products. Each statement adds one equation to the one already in the stem, which is enough. Note that the two statements lead to the same numbers; on the GMAT the statements never contradict each other. For the full method (AD/BCE elimination, yes/no questions, testing cases), see our guide to GMAT Data Sufficiency.
2. Table Analysis
A table similar to a spreadsheet, which you can sort by any column, followed by statements with two options each (Yes/No, True/False and similar). All statements must be judged correctly.
The table shows six stores of a retail chain last year.
Store Region Revenue ($ thousands) Staff Customer rating Alder North 420 12 4.1 Birch South 380 8 4.5 Cedar North 510 15 3.9 Dogwood East 290 7 4.3 Elm South 460 10 4.0 Fir East 350 9 4.4 For each statement, select Yes if it is true based on the table; otherwise select No.
- The store with the highest revenue per staff member is in the South region.
- The median revenue of the six stores is greater than $400 thousand.
- Every store with a customer rating of 4.3 or higher has fewer than 10 staff.
Statement 1: Yes. Revenue per staff member: Alder 35, Birch 47.5, Cedar 34, Dogwood about 41.4, Elm 46, Fir about 38.9 (in $ thousands). Birch, in the South, is highest. The calculator earns its place here.
Statement 2: No. Sort by revenue: 290, 350, 380, 420, 460, 510. With six values, the median is the average of the third and fourth: (380 + 420)/2 = 400. That is equal to $400 thousand, not greater. This is the trap: three stores are above 400, which feels like “more than half”.
Statement 3: Yes. Sort by rating: the stores rated 4.3 or higher are Birch (8 staff), Fir (9) and Dogwood (7). All have fewer than 10.
Answer: Yes, No, Yes. One wrong row and the whole question is wrong, so judge each statement against the exact condition it states (greater than, at least, every).
3. Graphics Interpretation
A graph or other image (mba.com lists scatter plots, x/y graphs, bar charts, pie charts and statistical distribution curves) with a statement to complete by choosing from drop-down menus.
A bar chart shows the number of units of a product sold in each quarter of last year: Q1, 1,200; Q2, 1,500; Q3, 1,800; Q4, 2,700.
Select from each drop-down menu the option that makes the statement most accurate.
Sales in Q4 were [ 50% / 80% / 125% / 150% ] higher than sales in Q2, and the percent increase in sales from Q3 to Q4 was [ greater than / equal to / less than ] the percent increase from Q1 to Q3.
First blank: 80%. (2,700 − 1,500)/1,500 = 1,200/1,500 = 0.8. The 125% option is the trap of comparing Q4 with Q1 instead of Q2: (2,700 − 1,200)/1,200 = 1.25.
Second blank: equal to. Q3 to Q4: 900/1,800 = 50%. Q1 to Q3: 600/1,200 = 50%. The bar for Q4 jumps visibly more, which pulls people towards “greater than”, but percent change depends on the starting value.
On the real test, read the axes, the units and any note under the chart before the statement; mba.com specifically warns about mismatches between the units in a graph and the units in the text.
4. Two-Part Analysis
A problem with two linked tasks and a single list of options. You choose one option for each of two columns; the same option can be correct for both. The problems can be quantitative, verbal or a mix: trade-offs, simultaneous equations, two elements of an argument.
An office manager spends exactly $6,000 on new laptops at $900 each and new monitors at $250 each, buying at least one of each.
In the table, select a number of laptops and a number of monitors that are consistent with this information. Make only one selection in each column.
Laptops Monitors Number ○ ○ 4 ○ ○ 5 ○ ○ 6 ○ ○ 8 ○ ○ 10 ○ ○ 12
Set up. With x laptops and y monitors, 900x + 250y = 6,000. Divide by 50: 18x + 5y = 120. Since 5y and 120 are multiples of 5, 18x must be too, so x is a multiple of 5. x = 5 gives 90 + 5y = 120, y = 6. x = 10 gives 180, already more than 120, so x = 5 is the only possibility.
Answer: Laptops 5, Monitors 6. Check: 5 × $900 + 6 × $250 = $4,500 + $1,500 = $6,000. The two tasks are dependent: the laptop number fixes the monitor number.
5. Multi-Source Reasoning
Information spread across several tabs (text, tables, charts), followed by several questions; the example below uses a set of Yes/No statements and a five-choice question. mba.com notes that questions may ask you to spot discrepancies between sources, draw inferences or decide which source is relevant.
Tab 1, Shipping policy. Orders placed before 2 p.m. on a weekday ship the same day. Orders placed at or after 2 p.m. on a weekday, or at any time on a weekend, ship on the next weekday. Express orders are delivered on the first weekday after shipping; standard orders on the third weekday after shipping.
Tab 2, Memo. Starting next month, the cut-off for same-day shipping will move from 2 p.m. to 4 p.m. for express orders only. Standard orders keep the 2 p.m. cut-off.
Tab 3, Orders last week.
Order Placed Service 101 Monday, 1:30 p.m. Standard 102 Tuesday, 3:15 p.m. Express 103 Friday, 5:00 p.m. Standard 104 Saturday, 10:00 a.m. Express
Question 1. For each statement, select Yes if it follows from the information provided; otherwise select No.
- Order 101 was delivered on Thursday.
- Under the policy that starts next month, an express order placed at the same day and time as order 102 would ship one day earlier than order 102 did.
- Orders 103 and 104 were delivered on the same day.
Statement 1: Yes. Placed Monday before 2 p.m., so it shipped Monday; the third weekday after Monday is Thursday.
Statement 2: Yes. Order 102 was placed after the 2 p.m. cut-off, so it shipped Wednesday. Under the new 4 p.m. cut-off for express orders, the same order would ship on Tuesday, one day earlier. This statement needs Tab 1, Tab 2 and Tab 3 together, which is typical of MSR.
Statement 3: No. Both shipped Monday (103 was placed after the Friday cut-off, 104 on a weekend). But 103 is standard, delivered Thursday, and 104 is express, delivered Tuesday. The trap is stopping at “both shipped Monday”.
Question 2. If order 103 had been an express order, on which day would it have been delivered?
(A) Saturday (B) Monday (C) Tuesday (D) Wednesday (E) Thursday
Answer: C. It was placed after the Friday cut-off, so it ships on Monday whatever the service; express means delivery on the first weekday after shipping, Tuesday. A assumes it shipped on Friday. The new 4 p.m. cut-off would not help either, since the order was placed at 5 p.m.
How to handle MSR: spend the first 60–90 seconds skimming each tab for what it contains (a rule, an exception, data), not for details. Then go back to the specific tab each question needs. The sources usually interact, as the memo does here, and the hardest questions sit exactly where one source modifies another.
Timing strategy
With 45 minutes for 20 questions, the average is 2 minutes 15 seconds, but the types differ a lot:
| Type | Rough target |
|---|---|
| Data Sufficiency | 1:45–2:00 |
| Table Analysis | 2:00–2:30 |
| Graphics Interpretation | 2:00–2:30 |
| Two-Part Analysis | 2:00–2:45 |
| Multi-Source Reasoning | 1:30 to read the tabs, then about 2:00 per question |
- Bank time on Data Sufficiency. DS questions are usually the quickest; use the time you save on MSR and the longer Two-Part Analysis questions.
- Check the clock at question 10. You should have at least 22 minutes left.
- Cap any question at about 3½ minutes. Make your best choice for every part, bookmark it and move on. Remember that one part wrong already means no credit, so polishing the last row of a question you are unsure about rarely pays.
- Leave nothing blank. If time is running out, enter answers for every remaining question.
How to practise Data Insights
- Learn the screens. Much of DI’s difficulty is the interface: sorting a table, switching tabs, reading two drop-downs. Practise on realistic, interactive screens rather than printed tables.
- Master Data Sufficiency first. It is the most distinctive format and appears throughout the section. Our Data Sufficiency guide has the method and practice questions.
- Revise the Quant you need for data. Percent change, ratios, averages and medians appear constantly in DI; see our guide to GMAT Quantitative Reasoning.
- Read the conditions word by word. “At least”, “greater than”, “every”, “only” decide most Table Analysis and MSR statements. Many misses are reading errors, not maths errors.
- Log misses by part. On multi-part questions, note which part you got wrong and why. Often it is the same kind of statement each time.
- Use official practice exams for the final check. GMAC’s free practice exams on mba.com include a full DI section with the real interface.
Common questions
Can I use a calculator in Data Insights? Yes. As of October 2026, mba.com states that an on-screen calculator is available in Data Insights. It is not available in Quant.
Is there partial credit in Data Insights? No. GMAC says that on multi-part questions you need every part right to score.
Can I go back to an earlier question? Only through Review & Edit. Per mba.com, once you have answered every question in the section and time remains, a review screen lists all questions and your bookmarks; you can open any of them and change up to three answers.
Is Data Insights harder than Quant? It is less familiar rather than harder: the maths is the same level, but the formats are unique to the GMAT. For the score side, see what a good GMAT score is.
To see every DI format next to Quant and Verbal questions, look at our GMAT sample questions. If you are deciding between tests, the GMAT vs GRE comparison and our GRE to GMAT conversion tool put the two scales side by side.
Also in Português (Brasil), Русский, Español