Command of evidence — quantitative

A graph or table, a short passage, and a question asking which choice uses the data to support a claim — or which claim the data supports.

The good news: these are the most reliably gettable hard-looking questions in the section, because the answer is a fact, not a judgement.

#The method, in order

1. Read the title of the figure. What is being measured, of what.
2. Read both axis labels, including the units. Percentages? Thousands? Per capita? Change from a baseline?
3. Read the legend. Which line or bar is which.
4. Only now read the question.
5. Find the specific data point(s) the choices refer to and check each against the figure.

Steps 1–3 take fifteen seconds and prevent nearly every miss on this question type. The wrong answers are almost all built from misreading the axis — quoting a raw number as a percentage, or reading the wrong series.

#What the wrong answers do

• Correct number, wrong series. The value is on the chart, just from the other line.
• Correct data, wrong claim. True about the figure, but doesn't support the claim the question named.
• Directionally reversed. "Increased" where the chart shows a fall.
• Comparison the data can't make. The claim compares two groups; the figure only shows one.
• Beyond the range. A statement about a year, category, or condition not in the figure. Whatever the trend looks like, the data doesn't extend there, so the claim isn't supported.

#The rule that resolves most of them

Supported means stated by the data, not consistent with the data. A choice can be entirely reasonable, fit the trend, and still be wrong because the figure doesn't show it. Ask of every choice: can I point at the exact number that makes this true? If you can't point at it, it isn't support.

#Speed

Once you have the habit, these take 60–75 seconds. The bulk of that is steps 1–3, and they're the ones under time pressure you'll be tempted to skip. Don't — skipping them is what turns a 60-second question into a 90-second wrong answer.