GMAT Verbal cheat sheet

Main idea = the claim the whole passage is organized to support; it must cover every paragraph, not just the first.

Primary purpose answers use a verb + object frame (e.g., 'explain a phenomenon', 'rebut a theory'), not a restatement of content.

Passage structure markers: 'however/but/yet' signal contrast (often the author's point); 'thus/therefore' signal conclusion; 'for example' signals support for a preceding claim.

Detail questions are answered only from stated text; a choice that is true in the real world but absent from the passage is wrong.

Inference answers must be necessarily true given the passage; 'likely' or 'could be' answers are wrong when the question demands certainty.

Tone/attitude is inferred from evaluative adjectives and adverbs (e.g., 'merely', 'remarkably'), not from topic alone.

When two answer choices seem correct, the wrong one typically overstates ('all', 'never'), narrows scope, or reverses a causal claim.

Author's viewpoint questions: distinguish the author from the researchers/theories the author describes; the author may report a view without endorsing it.

Argument structure: conclusion (the claim argued for) + premises (evidence offered). The conclusion is usually the statement the question asks you to weaken, strengthen, or evaluate.

Assumption = an unstated premise required for the conclusion to follow; negate it - if the argument collapses, it is a necessary assumption.

Weaken answers attack a link between premise and conclusion (alternative cause, reversed causation, unrepresentative sample); they need not disprove the conclusion.

Strengthen answers supply a missing link or rule out an alternative explanation; 'sufficient' strengthening is not required, only increased support.

Causal arguments (X caused Y) are weakened by: Y caused X, a third factor caused both, or the correlation was coincidental.

Sample-based arguments are weakened by evidence the sample is unrepresentative or too small; strengthened by evidence it is random and adequate.

Evaluate-the-argument questions are answered by a question whose two possible answers (yes/no) point in opposite directions for the conclusion.

Boldface questions: identify each bold statement's role - evidence, conclusion, or intermediate conclusion - and its relation to the main conclusion.

Subject-verb agreement: locate the true subject by stripping prepositional phrases and modifiers; the verb agrees with the head noun, not the nearest noun.

Pronoun reference: a pronoun must have exactly one logical antecedent that agrees in number and person; 'it/they' with ambiguous antecedents is wrong.

Modifier placement: an introductory modifying phrase must describe the subject of the main clause immediately following it (avoid dangling/misplaced modifiers).

Parallelism: items joined by 'and', 'or', 'but', 'not only...but also', or in a list must share grammatical form (noun with noun, clause with clause).

Comparison must be logically parallel and complete: compare like with like, and supply the missing verb or noun when ellipsis creates ambiguity.

Idiom and verb form: fixed prepositions and constructions (e.g., 'ability to', 'prohibit from', 'so ... that') are tested; unfamiliar idioms are usually not the deciding split.

Tense: use past perfect only for an action completed before another past action; do not shift tense without a time-based reason.

Countable vs. uncountable: 'fewer' with countable nouns, 'less' with uncountable; 'number of' with countable, 'amount of' with uncountable.

Split-and-compare strategy: find the first point of difference between answer choices, eliminate on a clear grammar rule, then move to the next difference.