Gre · Prep Guide
Read this first
Data Interpretation (DI) is not a separate question type — it is a set of standard Quant questions that all draw on the same presented data: a shared table and/or graph. This bible teaches the read-first method — decode the axis, scale, and legend, then estimate — and the named traps that make a wrong choice feel right. Work the practice set at the end as a mini-diagnostic.
The format on one page
| Question type | How you answer |
|---|---|
| Quantitative Comparison | Pick one of 4 fixed choices (A / B greater, equal, cannot be determined) |
| Select One | One answer from 5 choices |
| Select One or More | Select all that qualify — all-or-nothing, no partial credit |
| Numeric Entry | Type an integer/decimal in one box, or a fraction in two boxes |
| DI reuses the standard Quant formats — the data set is shared across the cluster. | |
One figure, a set of questions
Every Data Interpretation (DI) set opens with a shared figure: a table, a bar graph, a line graph, a circle (pie) graph, a scatterplot or boxplot, or a combination of these. The questions that follow all read from that one figure. Your job is to extract, estimate, and compare values from it accurately and quickly — the arithmetic is rarely the hard part; the reading is.
| Question format in a DI set | How you answer |
|---|---|
| Multiple-choice — Select One | Pick one answer from 5 choices. |
| Multiple-choice — Select One or More | Select all choices that qualify — no partial credit. |
| Numeric Entry | Type an integer or decimal in one box, or a fraction in two boxes (numerator / denominator). |
| Quantitative Comparison | Four fixed choices (A / B / C / D) comparing two quantities drawn from the figure. |
The graphs ARE drawn to scale
Therefore you may read, estimate, or compare values by sight or by measurement.
| Figure type | What it shows | Scale note to check first |
|---|---|---|
| Table | Exact values in rows and columns | Read the exact number; watch the units in the column header. |
| Bar graph | Category totals compared by bar height/length | Note whether the value axis starts at 0 or is broken. |
| Line graph | A quantity changing across an ordered axis (often time) | Check both axis scales; look for a dual axis. |
| Circle (pie) graph | Parts of a whole — the titled data equals 100% | Sector areas are proportional to their percents. |
| Scatterplot / boxplot | Spread, correlation, and five-number summaries | Drawn to scale; read positions directly. |
Read the figure before the questions
- Read the title and every label first. Title, axis labels, units, legend, and any footnote. This one pass tells you what the figure measures and in what units.
- Note the scale traps. Does the value axis start at 0? Is it broken? Is there a second axis? Are quantities absolute or percentages? Mark these before you read a single value.
- Estimate by sight where you can. The graphs are to scale, so a quick visual read often settles a comparison without arithmetic — the calculator is a last resort, not a first move.
- Match the format to the answer. Select-One wants one choice; Select-One-or-More wants every qualifier and no extras; Numeric Entry wants an exact value (round only if asked).
Figure (line graph): a store’s monthly revenue reads 200 thousand in March, 240 thousand in April, and 216 thousand in May.
Q1 (Numeric Entry): By what percent did revenue rise from March to April? (240 − 200) / 200 = 40⁄200 = 0.20 = 20%.
Q2 (Select One): April to May is a 10% decrease ((240 − 216) / 240 = 24⁄240 = 0.10). Is May revenue therefore back to the March figure, since it rose 20% then fell 10%? No. The 10% drop is taken on the larger April base: 240 × 0.90 = 216 ≠ 200. Percentages do not cancel unless they act on the same base.
Same figure, two formats, two different reads — and a distractor built on the assumption that +20% then −10% returns to the start.
Read the frame, then the questions
A Data Interpretation figure is not a picture to glance at; it is a labelled coordinate system. The GRE draws its bar, circle, and line graphs — and its coordinate systems — to scale, which means you are invited to read, estimate, and compare values by sight. But that invitation only pays off if you have first parsed the frame: what is on each axis, in what units, and on what scale.
| Frame element | What to extract before answering |
|---|---|
| Title | The single population the whole figure describes (e.g. “Households in Region X, 2019”). Every value inherits this scope. |
| Axis labels | What each axis measures. A common miss: reading the correct bar height off the wrong axis. |
| Units | Counts, dollars, percent, or per-thousand? The number on the axis is meaningless until you attach its unit. |
| Scale | Does the axis start at 0? Is it broken? Is there a second (right-hand) axis? Are ticks evenly spaced? |
| Legend / footnote | Which series is which, plus any “in thousands” or “excludes …” caveat that silently rescales every reading. |
- Read the frame. Title, axis labels, units, scale, legend, footnote — in that order. Do not look at the questions yet.
- Classify every value. For each number you will read, decide: is it a count, a percent, or a per-thousand rate? Write the unit next to it if it helps.
- Decide what the question wants. Percent-of-total (a share) and absolute change (a difference) are different questions. Underline which one is being asked.
- Estimate before you compute. The figure is to scale, so eyeball the answer first. Only reach for the basic on-screen calculator when the estimate does not separate the choices.
Frame: a bar graph titled “New Bicycles Sold, Store A, by Quarter (2022),” left axis labelled “Units sold,” ticks at 0, 200, 400, 600, 800. Bars: Q1 = 300, Q2 = 600, Q3 = 500, Q4 = 800.
Question: by how many units did sales change from Q1 to Q4?
Read the frame first: the axis is a count (units), and it starts at 0, so heights are directly comparable. Estimate: Q4 looks a bit taller than twice Q1.
Compute the absolute change: 800 − 300 = 500 units. That is a difference, not a percent — the stem asked “by how many units.”
Trap check: the tempting wrong move is (800−300)/300 × 100% ≈ 167% and reporting it as a count. Percent change answers a different question; here the answer is 500 units.
Frame: a table titled “Enrolment by Program, University Y, 2023.” Columns: Program, Students. Rows: Arts = 1,200, Science = 1,800, Business = 600, Other = 400.
Question: Science students are approximately what percent of total enrolment?
Classify the values: the column is a count. The question wants a percent of total, so you must build the base yourself.
Total base: 1,200 + 1,800 + 600 + 400 = 4,000. Share: 1,800⁄4,000 = 0.45 = 45%.
Estimate to sanity-check: Science is the biggest single row but under half the total, so a value just under 50% is right; 45% fits. Reading 1,800 against the wrong base (say, dividing by another single program) is the classic miss.
Frame: a line graph titled “Average Monthly Temperature, City Z (°C).” The vertical axis is broken and runs from 18 to 24 in steps of 2. The line reads Jun = 20, Jul = 22, Aug = 24.
Question: is August “twice as warm” as June, as the picture suggests?
Read the scale first: the axis does not start at 0 — it starts at 18. On a broken scale the Aug bar/point sits about three grid-steps above Jun and looks far taller, but the actual values are 24 vs. 20.
Compute honestly: the increase is 24 − 20 = 4°C, i.e. 4⁄20 × 100% = 20% warmer — nowhere near double.
Answer: no. A non-zero origin exaggerates vertical differences; always confirm the axis origin before you trust a “twice as tall” read by sight.
Read the frame before the number
Shared table (partial):
| Store | Share of chain sales | Chain total that month |
|---|---|---|
| North | 40% | $200,000 |
| South | 25% | $600,000 |
Question: Which store had the greater dollar sales that month?
Trap read: "North's 40% beats South's 25%, so North." That compares two rates and ignores that they sit on different bases.
Correct read: convert each share on its own base. North = 0.40 × 80,000. South = 0.25 × 150,000. South is greater — a smaller percent of a much larger base wins. The whole trap lives in reusing one base for both rows.
| The trap | What it rewards | The disconfirming check |
|---|---|---|
| Axis & units misread | Reading a value as-is — percent axis as a count, missing “in thousands,” a broken/dual axis | Say the axis first: label, units, multiplier, where it starts |
| Percent vs. absolute | Treating a rate and a count as the same currency | Multiply the rate by the correct base for that category before comparing |
| Wrong series / row | Right arithmetic on the wrong line, color, or row | Lock the legend/row key; confirm the row AND column that meet at the cell |
| Combining two figures | Chaining a share and a total on a mismatched key | Name the shared bridge quantity; confirm both figures use the same one |
| SD vs. range | Assuming a wider span means a larger standard deviation | Ask where the bulk sits about the mean, not how far the extremes reach |
GRE glossary
| Term | Meaning |
|---|---|
| Data Interpretation (DI) | A GRE Quant question type in which a set of questions all draw on the same data presentation — a table, graph, or chart — and ask you to read, combine, and reason about the values shown. |
| Data presentation | The shared table, bar graph, circle graph, line graph, or combination on which every question in a Data Interpretation set is based. |
| Drawn to scale | The convention that graphical data presentations (bar, circle, and line graphs) and coordinate systems are drawn accurately, so you may read, estimate, or compare values by sight or measurement. |
| Circle graph (pie chart) | A graph representing 100% of a titled data set, with each sector's area proportional to the percent it represents; reading a slice means reading its share of the whole. |
| Broken scale (scale break) | An axis on which part of the range is compressed or omitted, or that does not begin at 0, so bar heights and gaps can visually exaggerate differences unless you read the numbers. |
| Legend (key) | The chart annotation that maps colors, patterns, or line styles to categories; misreading or ignoring the legend is a common source of Data Interpretation errors. |
| Percent vs. count | The distinction between a share of a total (a percent) and an absolute number; a graph may show one while a question asks for the other, so the base must be identified before computing. |
| Numeric Entry | An answer format with no options in which you type the numeric answer into a box (or into a fraction's two boxes); common inside Data Interpretation sets alongside multiple-choice questions. |
| Select all that apply | A multiple-choice format that may have one or more correct answers, marked with square boxes; there is no partial credit, so every correct choice must be selected and no incorrect one. |
| On-screen calculator | The basic four-function-with-square-root calculator provided for the entire Quantitative Reasoning measure; it is not a graphing calculator and is best reserved for the multi-step arithmetic Data Interpretation often requires. |
| Section-level adaptive | The GRE design in which the difficulty of a measure's second Quant section is selected from your performance on the first; adaptation is per section, not per question. |
| Scaled score | The reported Quantitative Reasoning score from 130 to 170 in 1-point increments, reflecting both correct answers and the difficulty of the sections received. |
Frequently asked questions
What is a GRE Data Interpretation question?
Data Interpretation questions are a set of questions that all draw on the same data presentation — a table, a bar graph, a circle graph, a line graph, or a combination of them. Rather than testing one isolated calculation, the set asks you to read, combine, and reason about the values shown, applying percent, ratio, average, and probability ideas to the presented data.
Where do Data Interpretation questions appear on the GRE?
They are part of the Quantitative Reasoning measure, which has two sections (12 and 15 questions) for 27 scored Quant questions in total. Data Interpretation questions are grouped together because several questions share one graphic, but they are scored exactly like any other Quant question.
Are the graphs and tables in GRE Data Interpretation drawn to scale?
Yes. Graphical data presentations such as bar graphs, circle graphs, and line graphs are drawn to scale, so you may read, estimate, or compare data values by sight or by measurement. This is the opposite of bare geometric figures (lines, triangles, quadrilaterals), which are not necessarily to scale. Coordinate systems such as xy-planes and number lines are also drawn to scale.
What answer formats do Data Interpretation questions use?
A Data Interpretation set can mix formats: multiple-choice with one correct answer, multiple-choice with one or more correct answers (select all that apply), and Numeric Entry, where you type the number into a box rather than pick an option. Always read the instruction, because 'select all that apply' has no partial credit — you must mark every correct choice and no incorrect ones.
Can you use a calculator on GRE Data Interpretation?
Yes. A basic on-screen calculator (add, subtract, multiply, divide, square root) is provided for the entire Quantitative Reasoning measure, which includes Data Interpretation. It is not a graphing calculator and there is no Desmos. It is most useful for the multi-step arithmetic these sets often require, such as turning counts into percents.
How is GRE Quant scored?
The Quantitative Reasoning measure is scored on a scale of 130 to 170 in 1-point increments. Your score reflects both how many questions you answer correctly and the difficulty of the sections you receive, because Quant is section-level adaptive. There is no separate score for Data Interpretation on its own.
What is the single biggest mistake on Data Interpretation?
Reading a value straight off the plot without first reading the axis units, the scale, and the legend. Scales sometimes do not begin at 0, sometimes are broken, and a graph may show percents where a question asks for counts (or the reverse). Confirming what each number actually measures — and its base — before you compute prevents most Data Interpretation errors.
Should I guess on Data Interpretation?
Yes — always answer. GRE Quant has no penalty for a wrong answer, so a guess can only help. Because the questions in a set share one graphic, the reading work you do for the first question usually makes the others faster, so it pays to answer every question in the set.
Where to go from here
You now understand this GRE format better than most test-takers ever will. The points come from reps under the real timer, then from fixing the specific traps you keep falling for.
| Do this next | Why |
|---|---|
| Take an official ETS POWERPREP practice test | Convert format knowledge into reflexes under real timing. |
| Drill the other GRE question types | Verbal (TC, SE, RC) and Quant reward different reflexes. |
| Build a tiered vocabulary habit | GRE Verbal is vocabulary-defined — a little every day compounds. |
| Drill traps in the AskSia app | Per-distractor coaching on why you miss — bilingual, the part a static guide can’t give. |
