Make the text do the proving.
Main ideas, supporting evidence, logical inferences, and information in tables or graphs. Your job is to select what the supplied evidence supports, even when another choice sounds reasonable in real life.
Your step-by-step method
Name the task first
Read the question stem. “Main idea” asks for the passage’s overall point; “supports” asks for evidence for a particular claim; “most logically completes” often asks for a conclusion. This tells you what to look for when reading.
Build a one-sentence account
Identify who or what is being discussed, what the passage says about it, and any contrast or qualification. Treat words such as “however,” “some,” and “may” as meaning changes, not decoration.
Predict the shape of the answer
For a main idea, summarize the whole argument. For an inference, state the smallest conclusion the evidence justifies. For evidence, identify exactly what finding would strengthen or weaken the claim.
Check all parts of a choice
A choice can begin correctly and end with an unsupported explanation. Locate support for every important claim. Eliminate answers that add a cause, comparison, certainty, or population the passage never establishes.
Adjust for the question type
Central ideas
Separate the topic from the message. “Bird migration” is a topic; “tracking data challenged an earlier explanation of migration” is a main idea. An example in the passage often supports the message rather than being the message itself.
Evidence and inferences
Translate the claim into a testable relationship: if it is true, what should be higher, lower, present, or absent? Then compare the evidence to that prediction. An inference may connect two stated facts but should not require an invented third fact.
Graphs and tables
Read the title, axes, units, legend, and relevant row or group. Distinguish absolute amounts from percentages. If the question asks which data support a claim, the answer must both report the data accurately and support that exact claim.
See the strategy in action
Original BuildSAT examples that isolate the decision behind the method.
A study found that seedlings under blue light grew taller than seedlings under red light during one month. What conclusion is supported?
- Keep the comparison and scope: the measured outcome is height, for these seedlings over one month.
- “Blue light makes all plants healthier” adds a population and outcome that were not measured.
A claim says Treatment A improves survival relative to Treatment B. A table lists A: 72 survivors out of 90; B: 45 survivors out of 50. Does the table support the claim?
- The group sizes differ, so compare survival rates rather than counts.
- A: 72/90 = 80%. B: 45/50 = 90%. More survivors in A does not mean a higher survival rate.
Common traps to catch
- Choosing an answer because it repeats familiar words from the passage.
- Turning an observed association into proof of causation.
- Accepting “always” or “all” when the evidence supports only “sometimes” or a limited sample.
Can I point to the exact statement or data supporting every part of my answer?
Use your time well
On a return visit, reread the sentence relevant to the claim and its immediate context. Repeatedly rereading the entire passage without a specific question usually does not resolve a two-choice decision.