Critical Reasoning Inference Questions: Choosing What Must Be True
Inference questions reward the least exciting answer that cannot be false; ambition in an answer choice is a defect, not a virtue.
Why this matters
Inference and must-be-true questions invert the usual Critical Reasoning contract. Instead of evaluating an argument, you receive facts and must select the statement those facts guarantee. The trap is psychological: strong readers are trained to synthesize and extend, and the exam exploits that training by offering extensions that are plausible, interesting, and unguaranteed.
Scoring well requires recalibrating what correct means. On inference items the credited answer is often modest to the point of dullness, a restatement, a combination of two stated facts, or a mild logical consequence. Candidates who resist the pull of the interesting answer and embrace the airtight one convert this family from coin flips into near-certain points.
A working method
Read the stimulus as a fact inventory, noting quantifiers and qualifiers: some, most, all, never, typically. These small words carry the logical weight. Combine facts where they share terms, since many credited inferences arise from chaining two statements through a common element, such as some plus all yielding some.
Test each choice with the falsifiability question: can I construct a scenario consistent with every stated fact in which this choice is false. If yes, eliminate it regardless of plausibility. Be ruthless with comparatives, predictions, and causal claims the facts never made. Between two survivors, prefer the one with weaker wording; on this question type, could and some outlast will and most.
How to practice this skill
Drill counter-scenario construction explicitly: for ten inference items, attempt to defeat every answer choice, including the credited one. The credited answer should resist all attacks; documenting why builds the evidentiary standard this family demands. Candidates usually find the exercise slow for two sessions and automatic by the fifth.
Separately, drill quantifier arithmetic with short synthetic statements: what follows from most plus all, some plus none, or two overlapping most claims. Five minutes daily for a week covers the combinations the GMAT uses. The payoff is speed: recognizing a valid chain instantly leaves your time budget for the choices designed to tempt you.
A rigorous review protocol
Review misses by writing the exact sentence in the stimulus that the credited answer depends on, and the exact word in your chosen answer that overreached. The diagnosis is nearly always a single word, a stray all, a causal because, an unwarranted comparison. Naming the word trains the eye that timed conditions rely on.
Also audit your correct answers for process quality: did you eliminate on falsifiability or on vibes. A right answer chosen because it sounded safe is process debt. Rerun the counter-scenario test in review until the airtight standard, not intuition about tone, is what your eliminations actually execute.
Applying it in a timed section
Under time, spend your first thirty seconds on the fact inventory rather than jumping to choices; inference items front-load their difficulty in the stimulus. Eliminate aggressively on the first pass using single overreaching words, then apply the counter-scenario test only to the final two. If both survive honest attack, take the weaker claim and bank the time.
What mastery looks like
Mastery is a standard, not a technique: nothing gets your mark unless the facts guarantee it. When plausible-but-unproven answers start looking obviously defective, and modest restatements start looking obviously correct, you have internalized the burden of proof this family tests, and its questions become fast, repeatable points.