Command of Evidence - Quantitative

How to Approach It

Command of Evidence - Quantitative questions combine reading with data interpretation. You are given a claim and a table, graph, or chart. The correct answer must support the claim using the data. The biggest mistake is choosing a true data statement that does not match the claim. Treat the claim as a checklist. If the claim has two parts, the answer must support both. If it specifies a time period, the answer must use that period. If it says 'highest,' 'lowest,' 'largest increase,' or 'compared with,' the answer must make exactly that comparison.

Start by translating the claim into a data task. In the germination example, the claim says germination improves as soil moisture rises from very dry to moderately moist. That requires a trend across moisture levels, not just one data point. A choice saying 5% moisture has the lowest germination is true, but incomplete. A choice giving 20% moisture alone is also incomplete. The best answer cites the rise from 18% at 5% to 79% at 30%. It captures the direction and the range.

Two-variable claims are harder. If a historian says industrial output rose while agricultural employment fell, you need an answer that includes both variables. A data point about output alone is not enough. A data point about agricultural employment alone is not enough. The correct answer says output rose from 42 to 128 while agricultural employment fell from 51% to 24%. The SAT often includes true-but-partial choices to punish students who stop after seeing a familiar number. Do not stop at true. Ask whether the answer is complete.

For highest/lowest questions, verify the full column. If a researcher claims Downtown had the highest attendance, check every branch. The correct answer should say Downtown had 12,400 visitors, more than any other branch shown. A choice saying North had 8,900 visitors is true but irrelevant. A choice saying South exceeded West may be a valid comparison but not the comparison required. For combined-quality questions, such as high strength and low brittleness, avoid single-column traps. The best treatment may not have the absolute highest strength or the absolute lowest brittleness; it may have the best combination that matches the claim.

When stuck, annotate the graphic before reading choices. Circle the relevant columns and rows. Write a tiny note: full trend, highest value, two-part comparison, combined metric, or change over time. Then read each answer like a claim about the table. Is the number accurate? Is the row correct? Is the unit correct? Is the comparison the right one? Finally, check whether it supports the text claim. A mathematically accurate statement can still be wrong if it answers the wrong question. COE-Q rewards careful alignment: claim, data, comparison, and wording must all point in the same direction.

As you practice COE-Q, log the data-reading error precisely. Was it the wrong row, wrong column, wrong unit, wrong time period, partial claim, or irrelevant true fact? If the claim has two parts, mark whether your chosen answer supported only one. This makes quantitative evidence questions much more predictable.

A reliable COE-Q routine has four steps. First, underline the claim. Second, identify the required data operation: highest, lowest, increase, decrease, comparison, combined condition, or exception. Third, locate the relevant cells before reading the choices. Fourth, check that the answer's wording does not subtly change the claim. Many wrong choices are accurate descriptions of the table but not evidence for the sentence in the passage. In the treatment example, Treatment C has the highest strength, and Treatment D has the lowest brittleness, but the claim asks for the best combination of high strength and low brittleness. The correct answer is Treatment B because it balances both measures. In a change-over-time item, endpoints matter. A choice about only the final year may not prove growth unless it compares with the starting year. In a rate item, watch units: number of cases, percentage, rate per 1,000, and total count are not interchangeable. Whenever a table has multiple columns, expect at least one distractor to use the wrong column. Whenever it has multiple years, expect at least one distractor to use the wrong time span. Slow down at the chart; speed up only after the relevant cells are clear.

Practice Questions