What these questions look like
Central Ideas and Details questions sit in the Information and Ideas domain, and they come in two shapes. The first asks for the whole: "Which choice best states the main idea of the text?" The second asks for a piece: "Based on the text, which statement about X is true?" Each rides on one short mini-passage with a single question attached, and both are open-book. The answer is stated in the passage or provable directly from it. You are never asked to know anything from outside the box, and you are never rewarded for a reading the text does not support.
Scope is the whole game
Before you touch the choices, say the job out loud in your own words. A main-idea question wants the point of the entire passage. A detail question wants one specific fact. Get the scope wrong and you will happily pick a true statement that answers a different question.
Two failure shapes cause most of these misses. Too narrow: a real detail from the passage, correct on its own, offered as if it were the main point. It captures one sentence and ignores the rest. Too broad: a claim that sprawls past what the passage actually says, or is so general it would fit almost any text on the topic. The content is on-topic in both cases; the coverage is wrong. Naming the scope first, in words, is what lets you feel the mismatch. This is the same move that carries /guides/text-structure-purpose-questions, where too-narrow and too-broad are the house distractors.
Boring beats interesting
The correct main-idea answer usually restates what the passage literally said, often with a hedge still attached: may, suggests, one factor. It is not dramatic. It does not resolve the tension the passage left open. The interesting answers do all of that, which is exactly why they are wrong. A distractor is engineered to be more satisfying than the truth: it adds a cause the text never named, a value judgment the author never made, or a confident conclusion where the passage only raised a possibility.
That is why strong, sweeping language is a warning sign, not a selling point. An answer that says a practice is harmful, or that something should be banned, is making a claim the passage has to earn word for word. When the text only reports a concern, the answer that reports a concern wins. The test punishes certainty the passage did not license, a tell covered further in /guides/too-strong-answers.
The recycled-language trap
Here is the single most seductive wrong answer on these questions. It reuses exact words from the passage, sometimes a whole phrase, and bolts them to a claim the passage never made. Your eye lands on the familiar words, your gut says "match," and you pick it before you have checked whether the meaning matches too.
Matching words is not matching meaning. This is a true-but-irrelevant trap: every piece is real and in-text, but assembled to answer a question you were not asked (the pattern gets its own treatment in /guides/true-but-irrelevant-answers). It works by anchoring, the habit of letting the words on screen do your thinking instead of committing to a meaning first. The defense is to decide what the answer must say before you read the choices, so the recycled words have nothing to hook onto. That anti-anchoring move is the whole idea behind /guides/predict-then-peek.
A worked example
Read this the way the test wants you to, then predict before you look.
Rooftop beehives have multiplied across cities in the past decade. Supporters celebrate the trend for helping pollinators. But several ecologists warn that packing managed honeybees into a small area can crowd out native wild bees, which are often the better pollinators. The surge in city hives, they argue, may hide a quiet decline in the insects that matter most.
Say the point in one boring sentence: the rise of urban honeybee-keeping might be masking a drop in native wild bees. Now the choices.
A) Rooftop beehives have become common in cities over the past ten years. True, and lifted almost word for word from the passage, which is the tell. It is one opening detail, too narrow to be the main idea. Recycled language, no receipt for the whole point.
B) Urban beekeeping is a popular hobby that supports pollinators. This captures only the supporters, the half of the passage before the pivot. It ignores the "but" entirely. Partial coverage is still wrong coverage.
C) The growth of managed honeybee-keeping in cities may be obscuring a decline in native wild bees. Boring, hedged, and exactly what the last two sentences say. Put your finger on "may hide a quiet decline." There is the line. This is the answer.
D) Cities should ban rooftop beehives to protect native bees. Interesting, decisive, and unprovable. The passage reports a concern; it never prescribes a ban. Too strong, out of text, no line to point to.
Three of the four are true statements about the topic. Only one is provable as the main idea, and it is the least exciting one on the list.
The method: point to the line
One method handles this entire question type. Call it the point-to-the-line method, and it runs in two steps.
Step one, predict. Before you look at a single choice, say the main idea (or the specific detail) in one plain sentence of your own. Not clever, not complete: just the literal point. This is Predict-Then-Peek, and it exists to stop anchoring before it starts.
Step two, demand a receipt. For the choice you want to pick, put your finger on the exact line in the passage that proves it. Not a line that is near it, not a line that shares a word with it: a line whose meaning makes the answer true. If you cannot find that line, the answer is not yours, no matter how good it sounds. This is the Receipt Rule, the same discipline behind /guides/command-of-evidence and /guides/the-receipt-rule. A feeling is not a receipt, and a recycled word is not a receipt either.
The two steps kill both traps. Predicting first beats the recycled-language and interesting-but-unprovable answers, because you match against your own meaning, not the choices' vibe. Pointing to a line beats too-broad and too-narrow answers: a detail cannot produce a line that covers the whole passage, and an overreach cannot produce a line at all.
Run this self-test today. Pull your last practice section and find every Central Ideas or Details question you missed. For each one, try to underline the single line that proves the correct answer, and the single line that would have proved your wrong pick. Count how many of your wrong picks had no line at all. Those were not comprehension misses; they were anchoring, and they are wasted points you can close this week without learning any new content. If most of your misses do have a defensible line, your problem is reading the passage, not choosing the answer, and that is a different fix (start with /guides/inference-questions).
Practice it now: the free simplesat evidence drill runs Central Ideas and detail questions with the point-to-the-line receipt built into the feedback, so you see the exact proof line on every item. No signup, works as a guest: /?drill=command-of-evidence