Command of Evidence (Quantitative) — Best-Guess Strategies

How to Approach It

Hard COE-Q items punish students who read the table before they read the claim. The fix is to start with the claim, translate it into a specific data operation, and only then look at the table. When you are unsure, the right routine has four steps: classify the claim, isolate the data operation, find the relevant cells, and match the choice's wording exactly. The fourth step is what most students skip.

Step one is to classify the claim. There are about five claim shapes that show up regularly on COE-Q: (1) directional trend ('X improves as Y increases'); (2) two-variable change ('X rose while Y fell'); (3) single-entity comparison ('the downtown branch had the highest'); (4) combined-metric ('best combination of high X and low Y'); and (5) before-and-after change for two groups ('A grew while B did not'). Each shape requires a different data operation, and using the wrong operation produces the wrong answer even when you read the table correctly.

Step two is to translate the claim into a data operation. Directional trend requires endpoints across the range. Two-variable change requires both variables over the same time span. Single-entity comparison requires the named entity's value plus enough of the other values to show it leads. Combined-metric requires both metrics for the named entity. Before-and-after for two groups requires four numbers: each group's start and end. Write the operation down in shorthand: 'need endpoints,' 'need both vars across time,' 'need leader + others,' 'need both metrics,' 'need four numbers.' This shorthand prevents wrong-column and partial-claim traps.

Step three is to locate the relevant cells before reading choices. Circle the column the claim names. Circle the rows for the named entities. Note the time span. Cross out everything else; it is distractor fuel. With the relevant cells highlighted, the choices stop being four equally tempting statements; they become four statements you can quickly verify against your cells. A choice that uses uncircled cells is suspicious; a choice that names a cell not in the table is wrong on factual grounds.

Step four is the wording-match step. After identifying a likely correct choice, read it back against the claim word by word. Does the choice use the same variable the claim names? Does it cover the same time period? Does it name the same entity? Does it support both halves of a two-part claim? Many wrong answers are accurate data statements that drop or change one of these elements. This is the step that catches the silent partial-claim trap, which looks correct until you check whether 'while' or 'and' in the claim has a counterpart in the evidence.

When two choices remain and both look defensible, prefer the more complete one. COE-Q correct answers are typically the longest, most specific choice of the four. They name a variable, a number or trend, and either a comparison or a time span. Distractors often drop one of these elements to make space for irrelevant information. Length is not a guarantee, but completeness is: the choice that addresses every part of the claim wins, even if it feels more elaborate than the others.

When you must commit without certainty, default to the choice that mentions every variable in the claim. If the claim names two variables, the evidence should name two variables. If the claim names a time span, the evidence should name a time span. If the claim names a specific entity, the evidence should name that entity. The choice that covers everything in the claim is almost always right. The practice questions below model the routine on hard items where two choices look like plausible evidence at first.

Example Questions