Skip to content

SAT · technique

How to attack the reading section

On this page · 5 sections

On the digital SAT, "reading" means the Reading and Writing section. It is one section, and it always comes first. You get two modules of 27 questions, with 32 minutes for each, so 54 questions in 64 minutes. That works out to about 71 seconds per question. Each module includes 2 unscored pretest questions that you cannot identify, so treat every question as if it counts. There are no long passages. Every question comes with its own short text of roughly 25 to 150 words, and you answer exactly one question about it. The one exception is cross-text connections, which give you two short texts about the same topic and ask how they relate.

Each module runs in a fixed order of four domains. Craft and Structure comes first: words in context, text structure and purpose, and cross-text connections. Information and Ideas is next: central ideas, inferences, and evidence from a quotation, a table, or a graph. Standard English Conventions follows, which covers punctuation, sentence boundaries, verb forms, and agreement. Expression of Ideas closes the module with transitions and the bulleted-notes questions. Inside each domain, questions of the same type are grouped together, and each group runs from easiest to hardest. So difficulty drops back every time a new question type starts, not only at a new domain. For example, the words-in-context questions climb from easy to hard, and then text structure and purpose starts easy again.

The section is scored from 200 to 800, and that score is added to your Math score for a total of 400 to 1600. Wrong answers cost nothing, so a blank is always a wasted chance. The test is adaptive. Module 1 has a mix of easy, medium, and hard questions, and how you do there sends you to a harder or an easier Module 2. Only the harder Module 2 leads to the top of the scale. Your score also depends on which questions you got right, not only how many, because the scoring takes the difficulty of each question you answer into account. So Module 1 accuracy is worth more than its 27 questions suggest.

The skill being tested is narrow and learnable. Every answer is proved by the text on the screen, or by a grammar rule. Outside knowledge never decides a question, even when the passage is about an astronomer, an animal, or a museum you already know a lot about. Your job is to find the choice the text supports and to reject choices that sound smart but go further than the text does.

The steps

  1. 1

    1. Read the question line before the text

    Look at the question stem first, because it tells you how to read the text. A stem asking for the main purpose of the text means you read for the shape of the whole text. A stem asking for the most logical way to finish the text means you read for the chain of reasoning that leads to the blank. A stem asking which choice follows the rules of standard English means you need less of the text's argument, but you still need enough context to find the subject, the antecedent, and the time frame. Punctuation questions are mostly about structure, while tense and agreement questions need the surrounding sentences. A stem asking you to use the notes to reach a stated goal means the goal is the only thing that matters. Classifying the question takes two seconds and saves you from reading every text the same slow way.

  2. 2

    2. Read the text for its claim and its turn

    Read the whole text once, at normal speed, and ask what the author actually says. Then find the turn. Words like "however," "but," "yet," and "in reality" usually mark where the author's real point begins. Phrases like "often portrayed as" or "long assumed" signal the popular view the author is about to correct. A common pattern sets up a popular view and then corrects it. In that case the correction, not the popular view, is the main idea. Note who holds each view, too. In cross-text questions, keep Text 1's claim and Text 2's response as two separate notes in your head, because the question is always about how they connect.

  3. 3

    3. Predict the answer before you look at the choices

    Say a rough answer to yourself in plain words before you read A through D. For a purpose question, it might be "to explain why the old view was wrong." For a transition, it might be "this sentence is an example of the last one." For a words-in-context blank, use any simple word that fits, like "limited" or "surprising." Predicting matters because the wrong choices are written to sound reasonable once you see them. If you hold your own prediction first, you compare four choices against one idea instead of against each other.

  4. 4

    4. Eliminate with named reasons

    Kill wrong choices for a specific reason you could state out loud. Too broad: the text describes one case, but the choice says "always" or "all scientific revolutions." Wrong direction: the text says the model was equally accurate, but the choice says less accurate. Half right: the choice gets one part of the text and then adds a cause or claim the text never gives. Unsupported: the claim might be true in real life, but nothing in the text proves it. Too narrow: the choice describes one detail or example, not the whole text's point. In Bluebook, cross each wrong choice out on screen, so you can see what is left.

  5. 5

    5. For grammar and transitions, look at what changes

    Compare the four choices and find the one thing that differs. If only the punctuation differs, it is a Boundaries question. First check whether the blank joins two complete sentences. That join needs a period, a semicolon, a colon, a dash, or a comma plus a FANBOYS conjunction (for, and, nor, but, or, yet, so). A colon works only when the second clause explains or expands on the first. If the choices differ only by colon versus semicolon, check that relationship. A comma alone makes a comma splice, and no punctuation at all makes a run-on. If the blank does not join two sentences, check for a nonessential element that needs matching punctuation on both sides, a list, or a subject-verb split where no punctuation belongs. If the choices are "its" and "their," find the noun the pronoun points back to and check whether it is singular or plural. For transitions, read the sentence before the blank and the sentence after it, then name the relationship: example, contrast, result, or addition. Do not trust your ear here. A comma splice can sound fine when you read it quickly.

  6. 6

    6. For bulleted notes, match the goal and ignore the rest

    These questions give you a short list of notes (usually five or six bullet points) and a goal, such as "present both sides of the debate" or "explain the problem and present a solution." Underline every part of the goal. Then test each choice against each part. A choice that is accurate and well written still loses if it covers only one side or only the solution. You do not need to read the notes closely, because every answer choice is built from them and is usually accurate. The question is only which choice does the whole job the stem names.

  7. 7

    7. Run the clock by module, not by question

    You have about 71 seconds per question on average. Many conventions and transition questions take 30 seconds, which buys time for harder inference and cross-text questions. If a question passes about 90 seconds, pick your best remaining choice, flag it with Mark for Review, and move on. A guessed answer costs nothing, but a blank always scores zero. Meanwhile, one stuck question can cost you three easy ones at the end of the module. Aim to finish each module with three or four minutes left for flagged items. Never let the timer run out with a blank.

Worked through

Here is an original practice question, solved the way you should solve one under time.

Text: Ecologists long assumed that hummingbirds in mountain meadows choose flowers mainly by color, since many flowers they visit are red. In one field experiment, however, researchers set out red and white feeders filled with identical sugar water. Each day they swapped the colors at each spot. The birds kept returning to the spots where they had fed before, whatever color the feeder there was. These results suggest that ______

Question: Which choice is the most logical completion? A) hummingbirds cannot tell the difference between red and white. B) memory of location may matter more than color in how these hummingbirds find food. C) red flowers produce more nectar than white flowers do. D) these hummingbirds prefer white feeders to red ones.

Step 1, about 3 seconds. The stem asks for the most logical way to finish the text, so this is an inference question. You need to follow the reasoning chain to the blank.

Step 2, about 20 seconds. The text sets up an old view: birds pick by color. Then "however" marks the turn. The experiment kept the food identical, swapped the colors, and found the birds followed the spot, not the color. The old view is the setup, and the experiment is the point.

Step 3, about 5 seconds. Predict before reading the choices: "where they ate before matters more than color."

Step 4, about 25 seconds. Now eliminate with named reasons. A is unsupported. The birds ignored color, but that does not prove they cannot see it. A bird can see red and still choose by place. C is unsupported. The sugar water was identical in every feeder, so the text says nothing about which flowers have more nectar. D is unsupported. The birds showed no color preference at all. They followed location, so nothing in the text shows they prefer white. B matches the prediction. Notice its hedge: "may matter more." One experiment supports a cautious claim, and the right answer on inference questions is often the cautious one.

Answer: B, in under a minute. Two habits did the work. You found the "however" turn, so you knew the experiment was the point. You also predicted first, so the plausible-sounding A never pulled you in.

Common mistakes

  • Choosing an answer that is true in the real world but not proved by the text. If you need outside knowledge to defend a choice, it is wrong, even when the passage is about a topic you studied in class.

  • Picking the extreme version of a correct idea. Say a text describes how one astronomer slowly changed his mind over twenty years. A choice that says "scientific revolutions are always gradual" has gone further than that one case and loses.

  • Choosing a half-right answer: one that repeats a real detail and then adds a cause the text never gives. If a text says a bridge plan failed because of a faulty soil survey, a choice saying it failed because of the engineers' math errors names the real failure but the wrong cause.

  • On bulleted-notes questions, picking the most accurate or best-written sentence instead of the one that meets every part of the goal. If the goal is to present both sides of a debate over closing a downtown street to cars, a choice that gives only the shop owners' side fails.

  • Choosing a transition from the sentence after the blank alone. "In addition" and "For example" can both sound fine. You must check whether the new sentence adds a new point or illustrates the point already made.

  • Spending two or three minutes on one hard words-in-context question, then rushing or blanking the quick transition and notes questions at the end of the module. Since wrong answers cost nothing, a blank is the most expensive mistake on the test.

Checklist

  • I can name the four domains of a Reading and Writing module in order and say which question types sit in each.
  • I read the question stem before the text and can tell in a few seconds whether I am reading for purpose, inference, evidence, grammar, or a goal.
  • I can find the turn in a text, such as "however" or "in reality," and I know the author's point usually follows it.
  • I predict an answer in my own words before I read the four choices.
  • I can reject each wrong choice for a named reason: too broad, wrong direction, half right, unsupported, or too narrow.
  • On grammar questions, I compare the choices, find what changes, and check subjects, verbs, and antecedents instead of trusting my ear.
  • On bulleted-notes questions, I underline every part of the goal and check each choice against all of it.
  • I move on after about 90 seconds, flag the question, and finish each module with no blanks.

More SAT technique