The five shapes
Text structure and purpose passages are short, usually one paragraph, and they sit in the Craft and Structure domain. When the question asks how the text is organized or what its overall structure is, you are naming a shape. Nearly every passage is one of five:
- Claim, then evidence. A statement, then support for it.
- General, then specific. A broad idea narrowed to concrete instances.
- Problem, then solution. A problem named, then a fix proposed.
- Compare and contrast. Two things laid side by side, or set in opposition.
- Description, then reaction. Something described, then a response to it.
Here is a claim-then-evidence passage, invented:
Coastal wetlands blunt storm surge better than concrete seawalls do. In the Gulf towns that kept their marshes intact, flood damage after the last big hurricane ran far below that of neighboring towns that had paved their shorelines.
Shape: a claim (wetlands beat seawalls), then evidence (the three towns). Name that in three words and the structure question is basically answered before you meet the choices.
Purpose is the shape, said out loud
Main purpose questions ("What is the main purpose of the text?") are the same job one level up. The purpose is what the shape is for. If the shape is problem-then-solution, the purpose is "to propose a solution to X." If it is claim-then-evidence, the purpose is "to support the claim that Y." Name the shape, then say what that shape is doing.
Do not confuse purpose with main idea. The main idea is what the passage says, its actual point. The purpose is what the author is doing with it: arguing, describing, proposing, comparing. A passage's main idea might be "urban trees lower summer temperatures." Its purpose is "to present evidence for a claim about urban trees." One is content, the other is the job. For the pure content version of this skill, see central ideas questions; here, always answer with a verb.
Function questions: what one sentence is doing
The more common version of this question type underlines a single sentence and asks what it does: "The underlined sentence primarily serves to..." or "The author includes the underlined sentence mainly to...". These are function questions, and the trap built into every one is answering with content.
Function is not content. Two sentences with completely different words can do the identical job (both give an example). A choice can accurately summarize what a sentence says and still be wrong, because it names the wrong job. "What does it say" and "what does it do" are different questions, and this item type is built to punish students who answer the first when asked the second.
The menu of jobs is short. A sentence almost always does one of these:
- gives background or context
- states a claim
- gives evidence for a claim
- gives an example
- sets up what comes next
- qualifies or concedes a limit
- draws a conclusion
- compares or contrasts
- raises a counterpoint
Read the underlined sentence and pick the job before you look down. Invented example, with the underline marked:
Reef biologists once assumed warmer water would bleach every colony at the same rate. Not every reef responds this way; colder Pacific populations showed almost no change over the same decade.
The underlined sentence qualifies a claim: it concedes a limit on the "every colony" assumption. A choice reading "provides an example of coral bleaching" describes content in the same neighborhood but names the wrong job, so it loses. Say "qualifies the claim" first and that choice cannot get a grip on you. This is the same predict-first discipline behind words in context and transitions: commit before you peek.
Run the self-test today
Take ten structure, purpose, or function questions from any practice set. Before you read the choices, write in the margin:
- For whole-passage items, the shape in two or three words: claim → evidence, problem → solution, general → specific.
- For underlined items, the one-word job: evidence, example, qualifies, concludes, counterpoint.
Then reveal the choices and keep the one that matches your note. Grade your misses two ways. Named the shape or job wrong? That is a reading problem: slow down on the text. Named it right and still got talked out of it by a slick choice? That is anchoring, and the fix is trusting your prediction, the whole point of Predict-Then-Peek. Ten questions, one margin note each, and you know which half of the method is your leak.
The trap to name: wrong function
The wrong answers here cluster around one pattern: the wrong-function trap. The choice describes a real rhetorical job, just not this part's job. Learn its flavors and they stop working:
- Different part's job. The choice accurately describes what another sentence in the passage does, then pins it on the underlined one. Defense: keep firmly in mind which sentence is underlined, not just which jobs appear somewhere in the text.
- Mischaracterized. Names a job in the right neighborhood but sideways: calls an example an argument, a description a rebuttal.
- Wrong scope. Too broad (claims the sentence does more than it does) or too narrow (captures part of the job and misses the rest). The too-narrow choice is the hardest to reject because it "isn't wrong," just insufficient.
- Invented content. The most common distractor: it describes something the passage never states or implies, with no textual anchor at all. These overlap with plain out-of-text answers.
The tell that defeats every one of them is the same: you named the job in your own words first, so a confidently worded wrong choice has nothing to anchor to. For the full taxonomy of these traps, see SAT wrong answer types.
The method, in one move
Whether the stem underlines a sentence or asks about the whole text, the move is identical:
- Read the marked sentence, or the whole passage, and ask what it does, not what it says.
- Say the answer in your own words first: the shape, or the one-word job.
- Only then read the choices, and keep the one that matches your words. Reject any that invent content, attach a real job to the wrong sentence, or get the scope or direction wrong.
That is Predict-Then-Peek pointed at function instead of meaning, and the same muscle carries straight into rhetorical synthesis questions. It should be fast. If you are rereading a short passage three times to decide what it is doing, the problem is not the passage; it is a time-sink, and the cure is committing to a prediction instead of hunting for certainty.
Practice it now: the free simplesat command-of-evidence drill trains the exact muscle these questions test, reading a sentence for the job it does in an argument rather than the words it happens to use. Free, no signup, works as a guest.