GMAT · Guides · October 22, 2026 · 11 min read
GMAT Critical Reasoning: question types, how to attack each, traps and worked examples
Critical Reasoning (CR) is the GMAT Verbal question type built on a short argument, usually under 100 words, followed by one question with five answer choices. To answer it you find the argument’s conclusion, the evidence for it and the gap between the two, then pick the choice that does what the question asks: strengthen, weaken, fill the gap, explain a puzzle and so on. Below are the main question types, how to attack each, the wrong-answer traps the test relies on, and four worked examples written by us.
Where Critical Reasoning fits on the GMAT
Verbal has 23 questions in 45 minutes and only two question types: Critical Reasoning and Reading Comprehension. GMAC does not publish the split, but a typical section has about 9–10 CR questions and 4 reading passages. Sentence Correction is no longer on the exam. For the full structure, see the GMAT syllabus and exam pattern.
How to read an argument: conclusion, evidence, gap
Almost every CR question rewards the same three-step reading.
- Find the conclusion. It is the claim the author wants you to accept. Look for signal words such as so, therefore, concludes, plans to, will. In plan questions, the plan itself is the conclusion: “this action will achieve this goal”.
- Find the evidence. These are the facts offered in support. Treat them as true; CR never asks you to doubt a premise.
- Name the gap. Ask: what does the author take for granted in jumping from the evidence to the conclusion? Typical gaps are an ignored alternative cause, a sample that may not represent the group, a comparison between things that are not alike, or a step from “can” to “will”.
Say the gap in your own words before reading the choices; otherwise the most familiar-sounding choice tends to decide for you.
Tip: read the question stem before the argument. It tells you whether you are looking for a gap to attack (weaken), to close (strengthen, assumption) or to describe (flaw, boldface).
The question types and how to attack each
| Type | Typical stem | What the right answer does |
|---|---|---|
| Assumption | “an assumption on which the argument relies” | states something that must be true for the conclusion to follow |
| Strengthen | “most strengthens the argument” | makes the conclusion more likely, often by closing the gap |
| Weaken | “most seriously weakens” | makes the conclusion less likely, often with an alternative explanation |
| Evaluate | “most useful to determine in order to evaluate” | a question whose answer would change how good the argument is |
| Inference | “most strongly supported by the statements above” | must be true, or almost certainly true, based only on the text |
| Paradox | “most helps to explain the discrepancy” | lets two facts that seem to conflict both be true |
| Boldface | “the two portions in boldface play which roles” | describes the role of each part (conclusion, evidence, objection) |
| Flaw | “is flawed because it” | names the reasoning error in general terms |
| Complete | “most logically completes the argument” | supplies the premise or conclusion that the word because or since calls for |
| Plan | “most strongly indicates that the plan will succeed / fail” | shows whether the action really leads to the goal |
Assumption: use the negation test. Negate the choice (turn “some workers would” into “no workers would”). If the argument falls apart, it is the assumption; if it barely changes, it is not. Correct assumptions are often modest (“at least some”), while extreme ones (“no other”, “all”) usually fail.
Strengthen and weaken: work on the gap, not the topic. A strengthener rules out an alternative explanation or shows the evidence is representative; a weakener supplies the alternative cause or shows the sample is unusual.
Evaluate: think of two answers. Each choice is a question. If a “yes” would help the argument and a “no” would hurt it, that is the answer; if neither changes anything, eliminate it.
Inference: stay inside the text. The passage is a set of facts; the correct choice follows from them, often by combining two. Strong words such as always or most are warning signs.
Paradox: explain both sides. Write down the two facts that seem to clash. The answer gives a reason for the surprising one without contradicting the other. A choice that explains only half of the puzzle is the classic trap.
Boldface: label before you look. Mark each bold part as conclusion, evidence for, evidence against, or a view the author rejects, then match the labels to the choices.
Flaw: describe, do not fix. The answer names the gap in general terms, such as “treats a correlation as proof of cause”.
Complete: check the signal word. After because you need a supporting premise; after so, a conclusion.
Plan: compare action and goal. Look for a reason the action might not produce the goal (no one will use it, costs cancel the gain) and pick the option that confirms or removes it.
Common wrong-answer traps
Most wrong answers fall into a few repeatable patterns:
- Out of scope. The choice is about something the argument does not cover: costs when the argument is about yield, future plans when the evidence is about last year.
- Opposite. The choice does the reverse of what the question asks: it strengthens when you need to weaken. It is especially tempting because it is clearly relevant.
- Irrelevant comparison. The choice compares two groups or things, but the conclusion does not depend on that comparison (“summers in city A are hotter than in city B”).
- Too extreme. All, never, only, no other. More than the argument needs (in assumption questions) or more than the text supports (in inference questions).
- Shell game. The choice addresses a conclusion that is close to the real one but not the same: a claim about all customers when the argument is about new customers, or about profit when the argument is about sales.
- Half right. Common in paradox and boldface questions: the first half of the choice fits, the second half does not.
Four worked examples
All four are original questions from the Nomto bank, not official GMAT questions. Try each one before reading the analysis.
Example 1: Weaken (easy)
Last season, farmers in the Velden valley who switched to a new drought-tolerant variety of barley harvested 25 percent more grain per hectare than they had the season before. The regional agriculture office concludes that the new variety produces higher yields than the variety the farmers grew previously.
Which of the following, if true, most seriously weakens the argument?
(A) Seed for the new drought-tolerant variety costs Velden farmers about 30 percent more per hectare than seed for the variety they grew before.
(B) Last season was much wetter than the season before, and Velden farmers who kept the old variety also harvested far more grain per hectare.
(C) In trials at a research station near Velden, the new variety outyielded the old one in every dry year of a five-year study.
(D) Farmers in a neighboring valley, impressed by the reports from Velden, plan to try the new variety on part of their land next season.
(E) Some Velden farmers who grew the new variety found its grain harder to sell to brewers, who prefer the old variety for malting.
Conclusion: the new variety yields more. Evidence: farmers who switched harvested 25% more than the season before. Gap: something else that changed between the seasons could explain the jump.
Answer: B. A much wetter season, with old-variety farmers also harvesting far more, means the weather, not the seed, accounts for the rise. That is the alternative cause. C is the opposite trap: it supports the office. A and E are about cost and sales, not yield (out of scope), and D is about other farmers’ future plans.
Example 2: Assumption (negation test)
The owners of Fennick Gym plan to attract new members by extending the gym’s opening hours from 9 p.m. to midnight. They point out that many people in the neighborhood work evening shifts that end after 9 p.m. and so cannot use the gym under its current hours.
Which of the following is an assumption on which the argument relies?
(A) Most of Fennick Gym’s current members would rather the gym opened earlier in the morning than stayed open later at night on weekdays.
(B) Keeping the gym open until midnight will not raise its running costs for staff, lighting, or heating at all.
(C) People in the neighborhood who work evening shifts exercise more often, on average, than people who work during the day.
(D) Some of the neighborhood’s evening-shift workers would be interested in joining a gym that was open when they could visit it.
(E) No other gym within reach of the neighborhood is currently open after 9 p.m. on weekdays or on weekends.
Conclusion (the plan): later hours will bring in new members. Evidence: many neighbours finish work after 9 p.m. Gap: being unable to use the gym does not mean wanting to.
Answer: D. Negate it: none of the evening-shift workers has any interest in a gym. Then later hours attract no one, and the plan collapses. E is the too-extreme trap: negate it (“one other gym is open late”) and the plan still works, just with competition. B is about costs, which the plan does not depend on, and C is an irrelevant comparison with day workers.
Example 3: Evaluate
The city of Dunmere plans to plant 10,000 trees along its streets to keep the streets cooler in summer. Opponents say the plan will be a waste of money, because when the city of Arlow planted a similar number of street trees ten years ago, summer street temperatures in Arlow did not measurably fall.
Which of the following would it be most useful to determine in order to evaluate the opponents’ argument?
(A) Whether most of the trees planted in Arlow survived long enough to grow large
(B) Whether most of Dunmere’s residents support the city’s tree-planting plan
(C) Whether the trees in Arlow were planted by city workers or by private contractors
(D) Whether summers in Arlow are usually hotter than summers in Dunmere
(E) Whether the planting in Arlow was paid for by the city or by a private donor
Conclusion: planting will not cool Dunmere’s streets. Evidence: Arlow did the same and saw no cooling. Gap: Arlow is a fair test only if its trees actually grew into shade trees.
Answer: A. Try both answers. If most Arlow trees died young, Arlow never had the shade Dunmere plans for, and the comparison is weak. If they survived and grew, the opponents’ case is much stronger. Each answer changes the argument’s value, which is the test for evaluate. D is an irrelevant comparison: a difference in climate does not tell you whether trees cool streets.
Example 4: Paradox
When the Calder Park Zoo raised its standard adult admission price by 20 percent last spring, its managers expected attendance to fall, at least for a while. Instead, the zoo had more visitors over the following summer than in any previous summer since it opened.
Which of the following, if true, most helps to explain the unexpected result?
(A) Admission prices at the other major attractions in the region, such as museums and theme parks, also rose last spring.
(B) Before last spring, the zoo had raised its standard admission price only once in the previous ten years.
(C) Many visitors last summer complained that the food and drinks at the zoo’s cafe were overpriced.
(D) Attendance at the Calder Park Zoo is usually much higher in the summer months than in the winter months.
(E) Last spring, the zoo opened a large new elephant enclosure, which was widely covered in regional newspapers.
The puzzle: prices rose 20%, yet summer attendance hit a record. We need a reason for more people to come despite the higher price.
Answer: E. A major new attraction with wide press coverage gives people a new reason to visit, strong enough to outweigh the price rise. A is the half-right trap: higher prices elsewhere might explain why attendance did not fall, but not why it reached a record. D compares summer with winter, while the record is measured against earlier summers.
Timing: about two minutes per question
Verbal gives you 45 minutes for 23 questions, just under two minutes each. Reading Comprehension passages take longer to read at the start, so aim for about 2 minutes per CR question, and a bit less on the easier ones.
- First 30–40 seconds: read the stem, then the argument; find the conclusion, the evidence and the gap.
- Next minute: go through the choices once, eliminating by trap type. Do not stop to “prove” each wrong answer.
- Last 20–30 seconds: compare the two survivors against the gap you named.
- If you are past 3 minutes, pick the best remaining choice and move on. You can bookmark a question and change up to three answers per section on the Review & Edit screen if time remains, and guessing is always better than leaving a question unanswered, which is penalised.
To see CR alongside every other GMAT format, look at our GMAT sample questions. Logical Data Sufficiency in the Data Insights section uses many of the same skills; see the guide to GMAT Data Sufficiency. For timed practice with official questions, GMAC’s free practice exams are on mba.com.