What is GMAT Verbal Reasoning?
GMAT Verbal Reasoning is the reading-and-logic section of the GMAT Focus Edition: 23 multiple-choice questions in 45 minutes, drawn entirely from two question families — Reading Comprehension (RC) and Critical Reasoning (CR). It measures how precisely you can extract meaning from dense prose and evaluate the logic of short arguments.
The section has a reputation for being “subjective.” It isn't. Every question has exactly one defensible answer, and wrong choices are engineered around predictable weaknesses: going beyond the text, reversing logic, or being true-but- irrelevant. Verbal rewards test-takers who treat it as a precision game of evidence — find what the passage or argument actually supports, and refuse to choose anything more than that.
Format at a glance
| Feature | Detail |
|---|---|
Questions | 23 multiple-choice questions |
Time | 45 minutes (average ≈ 1 minute 57 seconds per question) |
Question families | Reading Comprehension and Critical Reasoning, mixed in roughly equal proportion |
Passages | About 3–4 RC passages of roughly 150–350 words, each with 3–4 attached questions |
Adaptive? | Yes — difficulty adjusts to your performance as you go |
Review & edit | Bookmark anything; change up to 3 answers per section if time remains |
Section score | 60–90 scale, weighted equally toward your 205–805 total |
What changed in the Focus Edition
| Change | Old GMAT | Current GMAT (Focus Edition) |
|---|---|---|
Sentence Correction | Roughly 40% of Verbal — grammar rules tested directly | Removed entirely. No grammar questions remain. |
Question count / time | 36 questions in 65 minutes | 23 questions in 45 minutes |
Skill profile | Comprehension + logic + grammar memorization | Pure comprehension and reasoning |
Preparation style | Rule-heavy drilling (idioms, agreement) | Active reading and argument analysis |
The removal of Sentence Correction is the single biggest verbal change in the exam's history. Non-native speakers often gain; native speakers who coasted on grammar intuition often lose. Either way, the preparation method is now identical for everyone: structured reading and disciplined argument evaluation.
Reading Comprehension — deep dive
RC passages come from business, science, history, and social science — deliberately dry, information-dense writing similar to The Economist or a research summary. Everything needed to answer is on the screen; outside knowledge is unnecessary and actively dangerous, because plausible-sounding facts that aren't in the passage are always wrong here.
Main idea & primary purpose
Why it matters: appears for nearly every passage — usually the first question
- The main idea is what the passage does, not just what it mentions: “challenges a common explanation” beats “discusses solar panels.”
- Wrong answers here are usually true details elevated to headline status — accurate but too narrow.
- Write a 5–8 word mental headline after each paragraph; the combination is your main idea.
Inference & supporting ideas
Why it matters: the highest-frequency detail-level question type
- Inference on the GMAT means “must be true given the text” — not “plausible” or “probably.”
- Correct inferences feel boringly safe: restatements, mild combinations of two lines, or obvious consequences.
- Wrong answers import outside knowledge or stretch quantifiers: the passage said some, the choice says most.
Logical structure & function
Why it matters: tests whether you read the skeleton, not just the flesh
- “Why does the author mention X?” — answers describe roles: to support a claim, concede a point, introduce a contrast, provide an example.
- Track signposts while reading: “however,” “in fact,” “for example,” “critics argue.” Function questions almost always target these joints.
Tone, attitude & application
Why it matters: smaller counts, but free points with the right habit
- Tone lives in word choice: “claims” signals skepticism; “demonstrates” signals endorsement.
- Application questions ask you to extend the passage's principle to a new scenario — match the underlying logic, not surface similarity.
The active-reading protocol
Why it matters: this single habit upgrades performance across every RC question type
- Read the first sentence slowly — passages typically state their agenda there.
- For each paragraph, form a 5–8 word summary in plain English.
- Note where the author signals attitude shifts and evidence.
- Never reread wholesale to answer; return only to the specific paragraph your mental map points to.
Critical Reasoning — deep dive
Each CR item presents an argument of two to four sentences plus one task. Before touching the choices, identify the three structural parts: conclusion (what the author wants you to believe), evidence (what they give you), and the gap between them. Every CR question type is a different operation performed on that gap.
Weaken
Why it matters: the most common CR type — master this first
- Correct answers introduce new information that breaks the evidence→conclusion link.
- The strongest weakens attack the causal link (“X causes Y” → maybe Y causes X, or Z causes both), not peripheral details.
- A weaken never needs to destroy the argument completely — making the conclusion less likely is enough.
Strengthen
Why it matters: mirror-image of Weaken — same skill, inverted sign
- Correct answers close the gap: rule out an alternative explanation, add supporting evidence, or confirm an assumption.
- Beware choices that strengthen a premise rather than the link — restating evidence feels helpful and does nothing.
Assumption
Why it matters: teaches the logic that powers every other type
- The assumption is the unstated premise the argument needs to hold.
- The negation test is definitive: negate the choice — if the argument collapses, it was an assumption; if nothing changes, it wasn't.
- Necessary assumptions (required) differ from sufficient ones (guarantee the conclusion). Read the stem closely for which is demanded.
Inference (CR flavor)
Why it matters: looks like RC inference but runs on argument logic
- Here you derive what must follow from the given statements — no outside facts allowed, ever.
- Correct answers combine two claims or restate one conservatively; dramatic conclusions are always wrong.
Evaluate & flaw questions
Why it matters: higher-difficulty items concentrated near the top of the adaptive range
- Evaluate: pick the question whose answer would most change your belief in the conclusion — usually a yes/no test over a hidden assumption (surveys representative? samples comparable?).
- Flaw: name the reasoning error — correlation mistaken for causation, biased sample, reversed conditionality, part-whole confusion.
A repeatable answer method
Both question families reward the same discipline. Run this loop on every item until it becomes automatic:
- Read the stem first — know your job before you read the content.
- Build the map — conclusion + evidence + gap (CR), or paragraph headlines (RC).
- Predict before looking — sketch what a right answer must contain.
- Eliminate ruthlessly — kill out-of-scope, too-extreme, reversed-logic, and half-right choices. Half-right is fully wrong.
- Match the survivor against your prediction — if none matches, your prediction was off, not the answers.
Pacing strategy
| Checkpoint | Elapsed time | Budget left | If you're behind |
|---|---|---|---|
Question 8 | ~16 min | 29 min for 15 | Cut second readings; trust first-pass maps |
Question 16 | ~31 min | 14 min for 7 | Cap RC questions at 2:30; move on |
Question 21 | ~40 min | 5 min for 2 | Answer within 60 seconds using pure elimination |
Question 23 | ≤45 min | — | Spend leftovers editing flagged answers (max 3) |
Note the asymmetry versus Quant: Verbal gives you less average time per question (~1:57 vs ~2:08) but shorter individual tasks. RC sets are the exception — the first question on a fresh passage may legitimately take 2:30 because the passage reading is amortized across 3–4 questions.
How Verbal scoring works
Verbal reports on the same 60–90 section scale, weighted equally into the 205–805 total. Its score distribution is notably compressed compared to Quant — historically, small scaled- score movements near the top produce large percentile swings. In practical terms: eliminating careless errors on medium questions moves your percentile faster than conquering rare hard ones.
A 12-week Verbal study plan
| Phase | Weeks | Daily focus | Exit criteria |
|---|---|---|---|
Reading engine | 1–3 | 20 min daily dense-article reading with paragraph summaries; learn the CR anatomy (conclusion/evidence/gap) | Can map any argument in under 30 seconds |
Type mastery | 4–6 | One CR type per week (Weaken → Strengthen → Assumption → Evaluate/Flaw); untimed accuracy focus | 75%+ per type untimed; can name why each wrong answer fails |
Integration | 7–9 | Mixed RC + CR timed blocks (10 questions); full error-log reviews | 70%+ mixed accuracy at ~1:55/question pace |
Simulation | 10–12 | Weekly full mocks; repair weakest two patterns between mocks | Stable target-section scores across consecutive mocks |
7 costly mistakes to avoid
| Mistake | Why it happens | The fix |
|---|---|---|
Drilling Sentence Correction | Legacy materials still circulate everywhere | Retire SC materials entirely; grammar rules are no longer tested |
Choosing “smart-sounding” answers | Academic register masquerades as correctness | Demand proof: point to the exact line or logical step |
Rereading passages wholesale | Anxiety about missing details | Trust paragraph maps; return surgically |
Importing outside knowledge | Familiar topics trigger autopilot | The passage is the entire universe — if it isn't stated or implied, it's wrong |
Ignoring quantifier swaps | Some/all, likely/always look identical at speed | Circle quantifiers in choices during review sessions to train the eye |
Skipping the prediction step | Feels slow; actually saves 20–30 seconds per item | Predict, then match — never shop among choices blind |
Untimed-only practice until late | Accuracy grows confidence that evaporates under clock | Mixed timed blocks begin no later than week 7 |
Frequently Asked Questions
Q1Is grammar tested on the current GMAT?
No. Sentence Correction was removed when the Focus Edition launched, so there are no direct grammar questions. Grammar matters only insofar as clear sentences are easier to read quickly.
Q2Is GMAT Verbal hard for non-native English speakers?
It's different now, and arguably fairer. Without Sentence Correction, the section tests comprehension and reasoning rather than idiom memorization. Many non-native test-takers find focused reading practice closes the gap faster than grammar drilling ever did.
Q3How many Reading Comprehension passages will I see?
Typically three to four passages of roughly 150–350 words, each carrying three to four questions. Passages arrive from business, science, and social science domains.
Q4Should I read the questions before the passage?
No for full passages — question stems reference line numbers or content you haven't seen, so pre-reading wastes time. Read the passage once, actively, then handle its questions as a set.
Q5Why did my accuracy drop when I added timing?
Timed conditions expose fragile skills. Accuracy drops are normal in weeks 4–7; the fix is maintaining your method (map, predict, eliminate) at higher speed, not abandoning the method.
Q6What's a good Verbal score?
Verbal scores cluster tightly near the top of the 60–90 scale. Because admissions committees read section scores separately, aim to keep Verbal within a few points of your other sections rather than maximizing one lopsidedly.
Q7How long should Verbal preparation take?
Eight to twelve weeks at about an hour per day covers most working professionals — reading habits take the longest to consolidate, so front-load the daily reading habit even before formal question practice begins.
Q8Are Verbal and Quant scores reported separately?
Yes. You receive a 60–90 score for each of the three sections plus the combined 205–805 total, along with percentile rankings for each.
Keep building the map
Pair Verbal with Quantitative Reasoning and Data Insights to cover every corner of the exam.