Inference questions reverse the polarity of the entire section. Everywhere else, you accept the conclusion’s existence and judge its support; here, there is no conclusion, the stimulus is a set of facts, and your job is to find the answer those facts prove. Not suggest. Not make likely. Prove. The single most expensive habit on this type is bringing the real world with you: the credited answer follows from the text alone, and outside knowledge is not a resource here, it is a contaminant.

What an Inference Question Actually Tests

Whether you can hold a small set of statements and extract only what they jointly entail. The skill is restraint: most wrong answers are reasonable, plausible, even probably true in real life. They lose because “probably true” is not the standard. Must be true is the standard, and it is merciless about quantifiers, the difference between “some,” “most,” and “all” decides nearly every hard question of the type.

How to Identify Them

Stems: “If the statements above are true, which one of the following must also be true?” or “can be properly inferred,” or “most strongly supported by the information above.” A note on that last one: “most strongly supported” softens the standard slightly, the credited answer may be a near-certainty rather than a strict entailment, but the discipline is identical: stay inside the text.

The Lovare Method, Four Steps

One: inventory the facts. No conclusion-hunting; just catalog what is established, flagging quantifiers and conditionals. Two: combine. Most credited inferences come from linking two statements, an overlap of groups, a chain of conditionals. Three: pre-phrase lightly. Note one or two combinations the facts license, but stay open; this type hides its answer in unexpected pairings. Four: prosecute each choice. The question for every answer is “where in the text is this proven?” If you cannot point to it, it isn’t the answer, however true it sounds.

A Worked Example (Original)

The Penrose Public Library reports: every book in its Rare Collection was donated by a private collector. Some books donated by private collectors are first editions. All first editions in the library are kept in the climate-controlled annex.

What must be true? (A) Some books in the Rare Collection are first editions. Tempting, and unproven. The first editions among donated books might all sit outside the Rare Collection. “Some donated books” overlaps “Rare Collection books” in source, not necessarily in membership. (B) If any book in the Rare Collection is a first edition, it is kept in the climate-controlled annex. This follows: all library first editions are in the annex, and Rare Collection books are library books. Conditional, modest, airtight, the house style of correct inference answers. The wrong answer was more interesting. The right answer was provable. That trade defines the type.

Quantifiers: Where the Points Actually Live

Three rules carry enormous weight. Some + all chains: “some A are B” plus “all B are C” yields “some A are C.” Most + most overlaps: two “most” statements about the same group guarantee an overlap, “most A are B” and “most A are C” proves some B are C. Some + some yields nothing: two “some” statements never combine. Drill these until they are reflexes; the test builds its hardest inference questions directly on top of them.

Withheld Tip: on “must be true” questions, run a falsifiability check on your finalist, try to invent a scenario consistent with every stimulus fact in which the answer is false. If you can build one, the answer fails, no matter how natural it sounds. Thirty seconds of adversarial imagination beats three re-reads.

The Traps, Named

The Reasonable Neighbor is true in the world and unproven in the text. The Overquantifier upgrades “some” to “most” or “most” to “all.” The Reversal flips a conditional, reading A→B as B→A. The Outside Expert imports your background knowledge and bills it as inference.

How to Train It: The Loop Applied

Inference misses split cleanly in blind review: if you correct them untimed, the clock is pushing you into plausibility-mode, train timed sections with the explicit rule “point to the proof.” If misses persist untimed, the quantifier mechanics themselves are the gap, and the Priority Stack should slot a logic-drills week before anything else. This is also the type most distorted by anxiety: under stress, the brain reaches for what feels familiar, exactly the Reasonable Neighbor trap. A widening Delta on inference questions late in sections is a stamina signal, not a knowledge one.

If You Only Do 3 Things

1. Drill the three quantifier combinations until producing them takes zero conscious effort.
2. Adopt “point to the proof” as a physical habit, finger on the supporting line before bubbling.
3. Run the falsifiability check on every finalist for two weeks of practice.

Inference Questions: Quick Answers

What’s the difference between “must be true” and “most strongly supported”?

Standard of proof. “Must be true” demands entailment; “most strongly supported” accepts a near-certainty the facts make overwhelmingly likely. Both forbid outside knowledge and reward the same textual discipline.

Why are correct answers so boring?

Because provable claims are modest claims. The test writes vivid wrong answers and quiet right ones on purpose, interestingness is bait, and conditional phrasing is usually the sound of the credited choice.

Should I diagram inference stimuli?

When conditionals or quantifiers stack, yes, a three-line diagram exposes combinations that prose hides. Pure-fact stimuli without logical operators usually don’t need it.

Related Playbooks

• LSAT Paradox Questions
• LSAT Sufficient Assumption Questions
• Logical Reasoning: The Complete Strategy Hub
• The 12-Week LSAT Study Schedule

Lovare’s Take

Inference questions train the most lawyerly habit on the exam: the separation of what is established from what is merely likely. Students who build that wall, text on one side, world on the other, find the type flips from treacherous to mechanical. The facts are all there. The discipline is staying inside them.