The GMAT Verbal Reasoning section is the part of the exam that tests how a candidate reads, reasons about, and edits dense English prose under timed conditions. On the classic GMAT it carried 36 questions across Reading Comprehension, Critical Reasoning, and Sentence Correction, and the section ran for 65 minutes. Under the GMAT Focus Edition, the structure was compressed: the Verbal section now holds roughly 23 questions in a 45-minute window, Sentence Correction was retired, and the remaining item families were reweighted. In both versions, the section contributes independently to a 60–90 scaled score and feeds the composite that admissions committees actually see. The rest of this article is a working anatomy: question families, the reasoning skill each one actually probes, scoring mechanics, and a preparation architecture built around the habits that move the scaled score rather than the habits that only feel productive.
The structural shape of the section on test day
Under the Focus Edition, Verbal sits as the second block of the exam after Quantitative, with a short optional break separating them. Candidates see a fixed number of questions, each presented one at a time, and they cannot return to a question once it has been submitted. There is no separately timed essay and no separately scored integrated reasoning buffer; the entire non-Quantitative portion of the test is Verbal. The adaptive engine chooses the first question at a difficulty calibrated to the candidate's pre-test self-reported profile, then routes the rest of the section based on whether the answer is right, how long the response took, and how the algorithm balances accuracy with pacing. A candidate who answers the first ten items with near-perfect accuracy will see a steady escalation in item difficulty, while a candidate who misses two of the first three will see the engine pull back to a lower band.
This routing matters more than most candidates realise. The 60–90 score is not a raw percentage: it is derived from the difficulty level of the questions you actually faced, the proportion you answered correctly, and the consistency of your performance across the section. In practice this means that finishing a question you can solve is worth more, scaled, than spending two minutes wrestling with a question the engine has not yet shown you. The section is short enough — about 23 items in 45 minutes — that one stuck question can consume the budget for two solvable questions. A workable target is to keep most items under roughly 110 seconds, with a small reserve bank for the hardest 3–4 passages and reasoning chains.
One other structural point: there is no experimental-section trick on the Verbal side anymore. Every Verbal question counts. Some candidates coming from older prep material still budget for a hidden unscored block; on the Focus Edition that block no longer exists in Verbal, and the time formerly lost to it has been folded back into the 45-minute main section.
How the section is laid out in the proctor screen
Each question appears centred, with the stem above and four or five answer choices below, labelled A through E. A timer in the upper corner counts down from 45:00. The candidate clicks a choice, then clicks confirm; after confirmation the question disappears and the next one loads. There is no scratchpad for the question itself, only the on-screen noteboard. A candidate cannot flag a question, cannot revisit it, and cannot change a confirmed answer. This is why the pacing habits discussed later in this article are not optional: the section is a one-pass examination, and the only way to keep options open is to keep moving.
Reading Comprehension: what the passage is actually doing
Reading Comprehension on the GMAT Verbal section is not a general reading-speed test. It is a controlled test of whether a candidate can extract an author's claim, a piece of evidence, an inference, and a structural function from a passage of roughly 200–350 words. Passages are drawn from business, social science, biological science, and occasionally a humanities topic such as literary theory or art history. The language is dense, the sentences are long, and the authorial voice is deliberately argumentative rather than narrative. About half of the Verbal section on the Focus Edition is Reading Comprehension, and the section carries questions in three families.
The first family is primary purpose. The candidate is asked to identify what the passage is mainly doing: arguing a thesis, contrasting two schools, describing a phenomenon and its mechanism, or refuting a common interpretation. The trap is the answer that is technically true about the passage but is too narrow to be the main purpose. A useful diagnostic for the test is to ask, before reading the choices: would I, in one sentence, describe the second paragraph's job? If you cannot, the primary purpose is probably not the answer the test wants.
The second family is detail and inference. Detail questions ask what the passage explicitly says; inference questions ask what the passage must imply. The common error is treating the two as the same. A detail question is satisfied by a sentence you can point to; an inference question is satisfied only by a claim that the passage logically forces. Phrases like "the author would most likely agree" and "it can be inferred" are inference markers, and they reward slower, more careful reading.
The third family is function and structure. These questions ask why a sentence is in the passage at all: what role it plays in the argument, what assumption it depends on, or what would weaken or strengthen the author's claim. A candidate who reads a passage only for content will struggle here, because function questions are about the architecture of the argument, not the facts.
Reading habits that move the score
The single highest-leverage habit is reading the first sentence of each paragraph before reading the body. Passages are built so that the topic sentence carries roughly 60–70% of the information the questions will test. A 30-second skim of topic sentences gives the candidate a thesis map; only then should the body sentences be read at full speed. The second habit is to mark, mentally or on the noteboard, the point at which the author switches from description to argument. Most wrong answers in the harder RC questions come from candidates who treat an argumentative claim as if it were a descriptive one, or vice versa.
Critical Reasoning: the reasoning engine inside the passage
Critical Reasoning items are short arguments, typically 50–120 words, followed by a question that targets a specific logical move: assumption, strengthen, weaken, inference, flaw, or evaluate. They are not Reading Comprehension and they are not logic games. They are testing whether the candidate can see the structure of an argument and predict which lever a question stem will pull. Roughly a third of the Verbal section under Focus is Critical Reasoning, and for many candidates this is the family that decides whether the scaled score climbs past the 80th percentile or stalls in the mid-range.
The argument in a CR item always has three parts: a conclusion, at least one piece of evidence, and an implicit or explicit assumption that links them. The conclusion is the claim the author wants the reader to accept. The evidence is what the author offers in support. The assumption is the unstated bridge. Most wrong answers come from candidates who read only the evidence, miss the conclusion, and pick an answer that sounds plausible against the evidence but does not actually address the link the question is asking about. A workable discipline is to write, in five words or fewer, what the conclusion is, before looking at the answer choices. If a candidate cannot articulate the conclusion, no amount of choice analysis will rescue the question.
Question stems cluster into six families, and each has a tell. Assumption stems ("which of the following is an assumption the argument depends on") reward answers that, if false, would collapse the argument. Strengthen stems reward answers that, if true, would make the conclusion more likely. Weaken stems reward the opposite. Flaw stems ask the candidate to name the gap in one or two words ("confuses correlation with cause", "relies on a sample that is not representative"). Inference stems ask what must be true given the argument, not what could be true. Evaluate stems ask what piece of information would help the reader judge the argument's strength.
Common pitfalls and how to avoid them
Three traps reappear on most CR drills. The first is the reverse-direction trap: the argument goes from evidence to conclusion, but the wrong answer describes a move from conclusion to evidence. The second is the scope-expansion trap: the wrong answer introduces a new entity, comparison, or population that the argument never mentioned. The third is the common-sense trap: an answer that is true in the real world but not actually supported by the passage. The defence against all three is to filter every answer choice against three filters: does it match the direction of the argument, does it stay inside the scope of the entities named, and does the passage actually license the claim. Most wrong answers fail at least one of these filters. Most right answers pass all three.
Sentence Correction was retired: what replaced it
The third Verbal question family on the classic GMAT was Sentence Correction: a single underlined sentence and five answer choices, with the candidate asked to choose the most grammatically and stylistically correct version. Under the GMAT Focus Edition, this family was removed. The reasoning behind the change is that Sentence Correction rewarded rule-recall as much as reasoning, and the test designers wanted Verbal to be a more uniform reading-and-reasoning section. The practical effect for a candidate is that grammar drilling, which used to occupy 30–40% of Verbal prep time, is now optional. The candidate who plans a 12-week schedule should reallocate those hours to Reading Comprehension and Critical Reasoning depth work rather than to idiomatic usage lists.
That said, two grammar habits still pay off, because the surviving Verbal items are written in dense English. The first is subject-verb agreement at long distance: a candidate who can scan a 35-word sentence and find the true subject, separated from the verb by three clauses, will avoid misreading the argument. The second is modifier placement: a misplaced modifier can flip the meaning of a passage, and Reading Comprehension items sometimes hinge on whether the candidate noticed that a phrase was meant to modify the noun two words earlier, not the one next to it.
Scoring mechanics: how the 60–90 scale is built
The Verbal section contributes an integer score from 60 to 90, in one-point increments, to the overall GMAT Focus composite, which also runs 60–90. The composite is not a simple average of the three sections: under the newer scoring report the percentile band is reported alongside the scaled score, and admissions committees are trained to read the percentile, not the raw number. A score of 84 and a score of 85 can sit in different percentile bands at different times of year because the percentile is calibrated against the rolling pool of recent test-takers. A candidate optimising for admissions should pick a percentile target for the schools on their shortlist, then reverse-engineer the scaled score required to land in that band.