Critical Reasoning is the section of the GMAT Focus Verbal where most candidates lose points without realising where the leak sits. A test-taker finishes a passage, eliminates two obviously wrong answers, picks between the remaining pair, marks B, and moves on. Two screens later the score report shows a Verbal 34, a CR sub-band under 60, and a feedback line that mentions argument evaluation. The problem is rarely raw reading ability. The problem is a small set of recurring error patterns that survive two passes, look defensible on first read, and only collapse under the third. This article walks through the seven traps I see most often in candidate log files, the exact sentence-level mechanism that creates them, and the 30-second fix that closes each one. The target audience is anyone preparing for the GMAT Focus who is stuck between V34 and V40, or who has plateaued above 80th percentile on Quant and cannot figure out why Verbal will not climb with it.
The 7 recurring CR error patterns, ranked by how often they reappear in candidate logs
Most GMAT Critical Reasoning wrong answers do not come from careless arithmetic or from missing a word in the passage. They come from answer choices that survive a shallow reading of the argument and only collapse once you stress-test the link between premise and conclusion. In my experience reviewing candidate log files, seven patterns account for the bulk of CR misses on the GMAT Focus. They appear in every Verbal mock and on test day. The list is not abstract. Each item is a sentence-level habit, and each one has a 30-second fix that does not require re-learning argument structure.
The seven patterns, in roughly the order I encounter them, are: stem-substitution, scope-creep, polarity-flip, paraphrase-trap, conclusion-projection, causation-confusion, and the defender bias. Some appear in every CR question type — strengthen, weaken, assumption, evaluate, inference, boldface, method, paradox, plan. Others cluster in specific stem families. Knowing which trap you tend to fall into, on which stem family, is worth more than a generic 'read the argument carefully' reminder. A candidate who can name their own top two patterns usually gains 3 to 5 CR points on the next mock without doing any new content review.
A useful diagnostic: pull the last 20 CR questions you missed, sort them by stem type, and count which pattern appears most often. Most candidates find that one or two patterns explain 60% of the misses. That is where to spend the next two weeks, not on a new prep book.
Stem-substitution: the most expensive single mistake on GMAT Critical Reasoning
Stem-substitution is what happens when the candidate reads the question stem, forms a mental model of what the answer should do, and then quietly swaps in a different task than the one actually printed. It is the single most expensive mistake I see in CR logs because it produces an answer that is logically fine but answers the wrong question. The candidate walks away thinking the test was unfair, when the test asked exactly what it said it would ask.
The classic version: the stem asks 'Which of the following would most strengthen the argument?' The candidate treats it as weaken, because the argument feels weak, and selects the answer that most damages the conclusion. Or the stem asks 'Which of the following must be true?' and the candidate treats it as a strengthen question, picking an answer that would make the argument more persuasive. Both answers can be coherent English sentences. Both can be factually correct in the abstract. Neither answers the printed task.
The 30-second fix is mechanical. Before reading the answer choices, paraphrase the stem aloud or in your head using a verb-noun template: verb (strengthen, weaken, assume, infer, evaluate, explain) + object (the argument, the conclusion, the plan, the discrepancy). Write the paraphrase down on the erasable notepad if you are in person, or in the scratch window if you are online. Then read the answer choices only through that lens. If a choice does something other than the paraphrased verb, it is wrong regardless of how smart it sounds.
Common pitfalls and how to avoid them
The pitfall on stem-substitution is over-confidence. Strong Verbal readers tend to skim the stem and trust their model of the question. On the GMAT Focus, where each CR question has roughly 90 to 110 seconds, that trust is misplaced. A second copy of candidates — call it the 30-second copy — benefits from a literal re-read of the stem after the first pass. The fix is to re-read the stem before clicking confirm, not before selecting an answer. Re-reading before selecting wastes time; re-reading before confirming takes three seconds and catches the swap.
Scope-creep: when the right answer reaches further than the argument actually supports
Scope-creep is the second most common pattern in CR logs. The argument supports a narrow claim; the candidate selects an answer that is true in a wider frame. On a strengthen question, this looks like picking a choice that would help a related but broader conclusion. On an inference question, it looks like selecting an answer that is consistent with the passage but not actually forced by it. On an assumption question, the trap answer is one that the argument needs only if the conclusion is interpreted more ambitiously than it is written.
The mechanism is always the same: the candidate notices that a piece of information is missing from the argument, assumes the missing piece must be the bridge, and picks the answer that fills the gap. The problem is that the missing piece is often not the bridge. The bridge is usually a narrower, less exciting claim that the candidate's eye skips because it does not look like it is doing work.
Consider a sample shape. Argument: a company finds that customer service calls drop 15% in months when a new self-service feature is heavily advertised, and concludes that the feature is responsible. Strengthen: a choice that says 'in months when the feature is not advertised, calls rise by 10%'. That is scope-creep. The argument is about the effect of advertising the feature. A choice about months when the feature is not advertised tests something else. The right strengthen is narrower: it must hold the advertising fixed and isolate the feature as the variable, or rule out a confound. Scope-creep answers are recognisable because they contain a key word from the passage but apply it in a context the passage never set up.
The 30-second fix is to underline the conclusion's exact scope and the premise's exact scope on the notepad. Then for each answer choice, ask: does this choice operate inside the same two scopes, or has one of them moved? If it has moved, the choice is scope-creep, and it is wrong even when it is interesting.
Polarity-flip: the answer that is technically true but points the wrong way
Polarity-flip is the trap where the answer choice is the right kind of object but its sign is reversed. A strengthen question receives a weaken-style answer. A weaken question receives a strengthen-style answer. A 'most likely to be true' inference receives an 'unlikely to be true' statement. The content is in the right neighbourhood; the direction is wrong.
Why this happens: the GMAT writes distractors by taking a near-correct idea and inverting the operative word. 'Increases' becomes 'decreases'. 'Most' becomes 'few'. 'Necessarily' becomes 'possibly'. The candidate, scanning at speed, registers the topic and the structure, and misses the inversion. On a 90-second budget, the inversion is easy to miss. On a 25-second re-read, it is obvious.
The 30-second fix is to train a polarising habit. For every answer choice on a CR question, isolate the single operative claim — usually a clause containing a comparative, a quantifier, or a modal — and read it twice. Many candidates skip this step because they trust the first read. The first read is exactly what the distractor is designed to exploit. The second read is what catches it.
Common pitfalls and how to avoid them
The pitfall on polarity-flip is treating it as a vocabulary test. It is not. It is a habit test. The candidate who trains the two-read habit on every answer choice will catch most polarity-flips within 10 to 15 hours of focused practice. The candidate who relies on vigilance alone will miss them intermittently for the entire prep cycle. Treat the two-read as a ritual, not a decision.
Paraphrase-trap: the answer that rephrases the argument instead of doing the work
Paraphrase-traps are answer choices that restate the argument's conclusion or premise in slightly different words, often with a quantifier or a connector swapped. They look relevant. They feel like they engage with the passage. They do no work at all. On strengthen, weaken, and assumption questions, paraphrase-traps are the most common 'plausible wrong' answer family. On inference questions, they appear as conclusions that are restatements of the passage but not strictly forced by it.
The mechanism: the candidate reads the argument, builds a summary, and then sees an answer choice that matches the summary. The cognitive shortcut 'this matches what I just read' fires, and the candidate selects it. The problem is that the task on a CR stem is rarely 'restate what you just read'. It is 'do something to what you just read'. The paraphrase-trap rewards the shortcut and punishes the task.
The 30-second fix is to ask, for every answer choice on strengthen, weaken, and assumption stems, 'What new piece of information does this introduce?' If the answer is none — if the choice is made entirely of words already in the passage — it is a paraphrase-trap and is wrong. A real strengthen introduces something the passage did not say. A real weaken introduces something the passage did not say. A real assumption introduces something the passage did not say but had to assume in order to reach the conclusion.
Conclusion-projection: when the answer choice borrows the conclusion's confidence but not its support
Conclusion-projection is the pattern where the answer choice makes a claim that is stronger, more general, or more confident than the conclusion in the passage. The argument concludes a modest claim — 'X is likely to be a contributing factor' — and the answer choice projects the same idea as 'X is the primary cause' or 'X is the only relevant factor'. The structure is similar; the magnitude is not.