Reading comprehension is the most expensive part of the verbal section in two currencies at once: it carries the largest share of verbal marks, and it eats the largest share of your clock. A candidate who reads every passage carefully from the first word to the last will usually run out of time, and running out of time is scored exactly the same as not knowing the answer.
The good news is that the passages are not testing whether you enjoy reading. They are testing a small set of operations: locating information, distinguishing the main point from a supporting detail, and following what the author does and does not claim. All three are trainable and all three are faster than careful full reading.
Change the order of operations: read the question first, then hunt through the passage.
Reading the questions first is not a shortcut for its own sake. It turns an undirected task into a search, and search is fast because you stop when you have found the thing you were told to find.
When the passage has several paragraphs, add one move: after the stems, skim only the first and last sentence of each paragraph and write a three-word label in the margin — what that paragraph does (setup, evidence, counter, result). You now know which paragraph each stem points at before you read deeply anywhere.
Six question types, two behaviours. Nearly every question on the paper falls into one of these, and each type tells you where in the passage to look.
| Type | What it asks | Where the answer is |
|---|---|---|
| Main idea | What is the passage mostly about? | First and last sentences, read together |
| Specific detail | A fact named in the question | The named lines, plus the sentence before and after |
| Inference | What follows, without being stated? | One step from the text — never a leap |
| Author's attitude | How does the author feel about it? | Adjectives and verbs of evaluation |
| Vocabulary in context | What does this word mean here? | Substitute each option into the sentence and re-read |
| Structure / function | Why is this paragraph here? | The paragraph's first sentence and its link to the previous one |
Answer the specific questions first, and the main-idea question last. Once you have answered three specific questions, you already know what the passage said, which makes the main-idea question cheap. Doing it first means guessing with less information and having to re-read later.
Inference has a hard boundary. The correct inference is the option that must be true given the text. Options that are merely plausible, or that are true in the real world but not argued in the passage, are wrong. The words that give it away are always, never, all, only — extreme options are very rarely the answer.
Vocabulary in context is a substitution test. Do not ask what the word usually means. Put each option into the sentence, read the whole sentence again, and keep the one that leaves the meaning of the sentence unchanged.
Scope discipline. An option that answers the question but goes beyond the passage is wrong. A favourite wrong answer is the correct idea attached to the wrong subject, or the correct subject pushed one step too far.
Multi-paragraph passages change where answers live, not what they are. The sample paper's long passages run three or four paragraphs, each with a job: one sets the scene, one gives the turn or conflict, one reports a result or consequence. The main-idea question is answered by the whole arc — first paragraph's opening and last paragraph's closing — not by any single paragraph. Detail questions stay local: they name a paragraph by quoting a line from it. Structure questions are the ones that force you across paragraphs: "The second paragraph primarily serves to…" means only "What does this paragraph do relative to the one before it?"
Consider this passage opening: Urban heat islands are usually blamed on traffic and air conditioning, and both do contribute. Yet measurements in three Pakistani cities suggest that the loss of tree cover accounts for a larger share of the rise than either source, a finding that has surprised the engineers who modelled the cities' energy demand.
A main-idea question here is answered by reading the first and last sentences together: the author qualifies a common explanation and points to a different dominant cause. The word Yet is the hinge, and the last clause tells you the finding was unexpected, which is why a question about the engineers' reaction also has an answer in the passage.
Now consider an inference option: "Tree planting would eliminate urban heat islands." The passage supports that tree loss contributes a large share; it says nothing about eliminating the effect, and the extreme verb eliminate goes beyond the text. That option is wrong for the most common reason in the whole section: scope.
A multi-paragraph worked path. A three-paragraph passage: P1 describes a canal built in the 1850s to move grain; P2 reports that the canal was abandoned after the railway arrived, though it still carries floodwater; P3 says historians now treat the canal as evidence that trade routes can outlive their original purpose.
Main idea, answered from the arc alone: P1's first sentence gives the original purpose, P3's last sentence gives the historians' reinterpretation — the passage is about a route whose role changed, not about canal engineering or railway speed. A detail stem that quotes "floodwater" sends you only to P2. A structure stem that asks why P2 is between P1 and P3 is answered by P2's own job: it supplies the event that creates the gap between original purpose and later reinterpretation. Each stem picks one paragraph, except the main idea, which picks the ends.
In the passage above, which sentence carries the author's main claim?
Which option is the classic wrong answer to an inference question?
Roughly how should time be divided across a passage with four questions?
In the three-paragraph canal passage, a stem asks why the second paragraph is placed between the first and the third. Where is the answer?
A multi-paragraph passage's main-idea question should be answered by reading:
Without looking back: which question type should be answered last, and why? And what test decides a vocabulary-in-context question?
The answer is: main idea, because the specific questions tell you what the passage contains; and substitution, putting each option into the sentence and keeping the one that leaves the meaning unchanged. On a multi-paragraph passage, main idea is still last — it is the one stem that needs the whole arc.
Fill in the blanks from memory:
The word that often marks the author's turn from the common view to their own is . Inference options are usually wrong when they contain extreme words such as . An option that goes beyond what the passage argues fails on . On a multi-paragraph passage, the main idea is built from the first paragraph's and the last paragraph's .
Take one practice passage and do it twice — then repeat the same drill on a multi-paragraph passage. The first time, read it fully and then answer the questions, and write down the seconds it took. The second time, on a fresh passage, read the question stems first, read only the first and last sentence of each paragraph, label each paragraph's job in three words, then hunt. Compare the two times and the two scores.
Most candidates find the second run is meaningfully faster and no less accurate, and the reason is worth naming: the first run spent time constructing a mental summary of a passage nobody asked about, while the second run read with a question already in hand.
On multi-paragraph passages the gain is larger, because the first run re-reads every paragraph for every stem, while the second run already knows which paragraph each stem points at. If the second run is faster but less accurate, the cause is nearly always one of two things. Either you read the two anchor sentences and then answered from memory instead of returning to the lines, or you accepted an option because it sounded true rather than because it was in the text. Both are fixed by returning to the passage for every answer, including the ones you feel sure about.
If the second run is not faster, check whether you actually read the stems first. The habit is easy to intend and easy to abandon under time pressure, and it only pays when it is done consistently.
The verbal module is not finished yet. The next lesson sharpens the same reading skills — tone, purpose and inference — and shows how the question's wording tells you which kind of reading it wants before you answer it.