GRE Data Interpretation: Reading the Graph Before the Question

A data interpretation set presents one or more graphs, tables, or charts, followed by several questions drawn from the same display. The arithmetic involved is usually modest. What the set tests is whether you read the display correctly before computing.

That distinction explains most missed questions here. Students who miss them rarely miscalculate; they answer a slightly different question than the one asked, because they misread a unit, an axis label, or which of two similar categories the question named.

The strategic implication is that time spent orienting to the display before reading any question is not overhead — it is the work. Because several questions share one display, that orientation cost is amortized across the whole set.

Before reading question one, spend a moment establishing four things.

What each axis represents, and in what units. Thousands, millions, and percentages are all common, and a y-axis labeled "revenue (in thousands)" will produce answers that are off by a factor of a thousand if you miss it.

What the categories are, and whether any two are similar enough to confuse. Displays frequently include closely related categories — "total employees" and "full-time employees," or "applications" and "admissions" — precisely because questions can then hinge on which one you used.

Whether there is a footnote or a note beneath the chart. These are easy to skip and are usually load-bearing. A note saying figures exclude a particular region, or that percentages do not sum to 100 because respondents could select multiple answers, is there because a question depends on it.

Whether the scale is linear, whether it starts at zero, and whether a break appears in the axis. A truncated axis makes differences look larger than they are, which matters if you are estimating visually.

Most data interpretation sets include at least one percent question, and these generate the most errors because several distinct quantities sound alike.

The last row is the one worth memorizing. Percentage points and percent change are different quantities, and questions that involve a percentage on the y-axis are frequently built around the difference. Both values will appear among the answer choices.

A second recurring trap: when a question asks for the percent increase over several periods, you cannot add the individual percent increases. Successive percentage changes compound, so a 10% rise followed by a 10% rise is a 21% total increase, not 20%.

Look at the answer choices before computing. When they are far apart, estimation is faster and carries little risk. When they are close together, or when two differ by a factor that matches a common misreading, compute precisely.

Estimation is particularly effective on questions asking which category is largest, whether a ratio exceeds some threshold, or roughly what fraction of a total a segment represents. Reading a bar to the nearest gridline is usually enough.

It is least safe when the display has a truncated axis, since visual comparison of bar heights becomes misleading, or when the question involves a difference between two similar values, where small reading errors become large proportional ones.

One useful check regardless of method: ask whether your answer is plausible given the display. An answer implying a category grew tenfold when the bars look similar is worth a second look before you move on.

Using the wrong denominator on a percent-increase question is the most frequent error, and it happens because the question's phrasing puts the two values in the opposite order from the formula. "How much greater was 2024 than 2019" mentions 2024 first, but 2019 is the denominator.

The second is answering from the wrong series when a chart has multiple lines or stacked bars. Stacked bars are especially error-prone, because reading a segment's value requires subtracting the boundary below it rather than reading its top edge against the axis.

Third is ignoring the note beneath the display. If a footnote exists, at least one question in the set almost certainly depends on it.

Fourth is spending too long on one question in a set. Since the display is shared, the orientation work carries over — but an individual question can still be disproportionately time-consuming, and the remaining questions in the set may be quicker.

Verbloom's GRE practice includes quantitative reasoning with explanations that show where a misreading rather than a miscalculation produced the wrong answer.

For data interpretation specifically, reviewing which part of the display you misread is more useful than redoing the arithmetic, since that is where the points are lost.