What is a Data Interpretation set?
In Quantitative Reasoning, ETS groups some questions into sets: all refer to the same table, chart, or other data display and ask you to interpret or analyze that data. Questions in a set may be multiple-choice with one or more answers, or numeric-entry. The question format can change; the data set stays the same.
Do not read too much into one point: the cited ETS pages do not specify the number of sets per section or questions per set. They state that a question is either independent or part of a set. The published structure has two Quant sections: twelve questions in twenty-one minutes, followed by fifteen in twenty-six minutes. For the shorter GRE, introduced on September 22, 2023, ETS states in its 2023 FAQs that the test covers the same content as before, and its Quant page still describes these sets. Prepare for the format, not a set quota.
What does a chart drawn to scale allow you to do?
ETS distinguishes between two kinds of figures. Geometry figures—lines, circles, triangles, and quadrilaterals—are not necessarily drawn to scale: you cannot judge a length or angle by eye. Coordinate systems and data displays—bar charts, pie charts, and curves—are drawn to scale, so you can read, estimate, or compare values by sight or measurement. For a data chart, visual inspection is allowed, unlike for geometry.
This guarantee applies to the scale shown, not necessarily to zero. ETS’s sample-question guide also warns about broken scales and bars that do not start at 0. Using this article’s data: if the Student visitor bars, 60 and 48 thousand, are plotted on an axis that starts at 40, they are 20 and 8 units high. The second appears two and a half times shorter than the first, although it is 80% of its value. Compare bar heights directly only when the axis is linear, starts at 0, and has no break; otherwise, read values from the scale.
In what order should you read a data set?
Scan the display instead of studying it cell by cell: ETS recommends identifying what it is about, then focusing only on what the questions ask. The protocol has seven steps (listed below), from the quickest to the most involved.
One rule applies to all these steps: answer using only the data, common facts (such as the number of days in a year), and your mathematical knowledge. Information you remember about the real-world context—a museum or a market—that cannot be derived from the data does not belong in your answer.
- Title: what the data set is about, the population, and the time period.
- Units: thousands, millions, counts, or percentages. ETS recommends checking the titles, labels, and legends.
- Scale and legend: intervals, where the axis starts, what the colors or patterns represent, and any notes clarifying the data.
- Question: reread what is being asked—a count, a share, a relative or absolute change—and the expected answer format.
- Targeted reading: note only the relevant cells or bars on your scratch paper, with their units.
- Estimate before calculating: round the data, predict the approximate value and direction of the answer. This often lets you eliminate choices before calculating.
- Plausibility check: correct unit, correct sign, and a value consistent with your estimate.
A fictional data set
The following three exercises use the same data set, created for this article: visitors to a fictional museum, in thousands, by ticket category. Apply the first steps before reading a question: check the title, units, and the fact that no total row is provided. Calculate the total when a question asks for a share or a change in the total number of visitors.
| Ticket category | Visitors in 2023 (thousands) | Visitors in 2024 (thousands) |
|---|---|---|
| Adult | 120 | 140 |
| Student | 60 | 48 |
| Senior | 40 | 44 |
| Child | 30 | 36 |
| Group | 50 | 52 |
Exercise 1: a count is not a rate
Single-answer question. Before dividing, express each change as a fraction of its own base: you can often identify the largest fraction without calculating.
Refer to the table above. Which category had the greatest percent increase in visitors from 2023 to 2024?
Choose an answer.
The solution, step by step
D: Child.
The method in action
- Each change is measured against its 2023 base. Adult: 20 out of 120, or 1/6 (16.7%). Senior: 4 out of 40, or 10%. Child: 6 out of 30, or 1/5 (20%). Group: 2 out of 50, or 4%. Since 1/5 is greater than 1/6, Child has the largest increase; no calculator division is needed.
- A is the count trap: Adult gained the most visitors in absolute terms (20 thousand), but it also had the largest base. B is a 20% decrease, so it is negative. C and E increased by less than Child in both relative and absolute terms.
Exercise 2: shares, totals, and bases
Multiple-answer question: when no number is specified, select every true statement and no others. First calculate both totals, 300 and then 320; several statements depend on using the correct total or base.
Refer to the table above. Which of the following statements are true? Select all that apply.
Select all correct answers.
The solution, step by step
A and B.
The method in action
- A is true: 140 out of 320 is 43.75%. Check: 40% of 320 is 128, and 140 is higher. B is true: Group increases from 50 to 52, but its share falls from 50/300 (16.7%) to 52/320 (16.25%). To maintain its share, Group would have needed to reach 53.3 thousand. A category can therefore grow while making up a smaller share if the total grows faster.
- C is false: the decrease is 12 out of a base of 60, exactly 20%, not more. Dividing 12 by 48 would give 25%—the wrong base, which is the ending value. D is false: 10% of 300 is 30, and the total grew by only 20 thousand, or 6.7%. Adult, Senior, Child, and Group gained 32 thousand visitors in total, but the 12 thousand loss for Student brings the net increase to 20 thousand. E is false: 30/300 is 10% in 2023, while 36/320 is 11.25% in 2024; using 300 as the total would give 12%, another total error.
Exercise 3: project a relative change
Numeric-entry question, with no answer choices to guide you. You must use the requested unit—thousands—and choose the right base: the 2024 value, not the 2023 value.
Refer to the table above. Student visitors decreased from 2023 to 2024. Suppose the number of Student visitors decreases by the same percent from 2024 to 2025. How many Student visitors, in thousands, will there be in 2025? Enter a number.
Enter your answer before viewing the review.
The solution, step by step
38.4.
The method in action
- The decrease from 2023 to 2024 is 12 out of 60, or 20%. Applied to 48, it leaves 80% of 48, which is 38.4; subtracting 20% of 48, or 9.6, gives the same result. ETS requires a period as the decimal separator for entry: 38.4.
- 36 results from repeating the same absolute decrease of 12, but the prompt asks for the same percentage decrease on a smaller base. 38,400 ignores the requested unit, thousands. Estimate: since the base is smaller, the decrease should be less than 12, so the result must fall between 36 and 48.
The pitfall
Reading a bar by eye or reusing the previous question’s total
Answering by eye or habit means reading a bar without checking where the axis starts, or reusing one question’s total for the next. The data set stays the same, but the base changes from question to question. The 2023 total is not the 2024 total, and the base for a change is the starting value, not the ending value.
The opposite extreme wastes time: copying the entire display before you know what you are looking for. ETS recommends scanning it first, then studying only what the questions require in detail.
How can you check an answer before submitting it?
A few quick checks catch the errors described above. Shares for categories that make up the same total without overlap add up to 100%, allowing for rounding: 40, 20, about 13.3, 10, and 16.7% for 2023. A large discrepancy is a reason to check the total and categories used. A decrease gives a result below the starting value; an increase gives a higher one. A change in share is expressed in percentage points: Student’s share falls from 20% to 15%, a decrease of 5 points, or one-quarter in relative terms. Finally, answer in the unit requested: 38.4 if the question asks for thousands.
The on-screen calculator provided in Quant performs the operations you choose but does not check the base you used. In Numeric Entry, the button that transfers the calculator display to the answer box does not work for a fraction, and ETS asks for the exact value unless rounding is requested. There is no penalty for a wrong answer, so answer every question.
Key takeaways
- Read the title, units, scale, and legend before the first question without studying the entire display.
- A data chart is drawn to scale and should be read using its intervals after checking where the axis starts; a geometry figure may not be drawn to scale.
- For each calculation, identify the base: the correct total, starting value, and unit.
- Estimate before calculating, then check the sign, unit, and approximate value.