Why Inference Questions Reward the Timid Reader

Most Logical Reasoning questions ask you to do something to an argument — weaken it, find its assumption, name its flaw. Inference questions are different. They don't give you an argument to attack. They give you a set of facts and ask: what can you prove from these facts alone? The correct answer is the one that must be true — or, in the softer variant, the one that is most strongly supported.

Here's the counterintuitive part: on inference questions, the boldest, most interesting-sounding answer is almost always wrong. The test rewards the reader who refuses to add anything. If the stimulus says 'most' and an answer says 'all,' that answer is dead — no matter how sensible it seems. Your job is to be relentlessly literal.

These are original questions written in the LSAT style. They are not reproduced from any real LSAT — LSAC's actual questions are copyrighted, and you should drill those on LawHub, LSAC's official platform. Use the set below to build the habit; use LawHub to test it under real conditions.

The Method: A 4-Step Attack for Inference Questions

1. Read the stimulus as a set of facts, not an argument. There's no conclusion to judge. Note every claim and — critically — the exact strength of each (some / most / all / only / if-then).
2. Do NOT prephrase a specific answer. Unlike assumption or strengthen questions, inference stimuli can support many valid conclusions. Trying to predict the one they'll ask about usually fails. Instead, hold the facts in mind and go to the choices.
3. Test each choice against the text. Ask: 'Does the passage guarantee this?' For a 'must be true' question, one counterexample kills a choice. For a 'most supported' question, ask which choice the evidence pushes toward hardest.
4. Eliminate anything that adds information. New terms, stronger quantifiers, causal claims the text never made, predictions about the future — all disqualifying. The right answer is often the most cautious, boring statement on the page.

The Traps to Expect

• Too strong. The stimulus supports a modest claim; the answer overstates it ('always,' 'never,' 'must,' 'cannot' where the text said 'often' or 'usually').
• Out of scope. The answer introduces a comparison, a cause, or a category the stimulus never mentioned.
• Reversed logic. The stimulus gives you 'If A then B'; the answer asserts 'If B then A' or 'If not A then not B' — neither of which follows.
• Half-right. The first clause matches the text perfectly, and the second clause quietly adds something unsupported.
• Real-world plausible. The answer is a true statement about the world that the stimulus doesn't actually prove. Truth isn't the test — provability from the passage is.

How to Read Quantifiers (Where Inference Credit Lives)

Inference questions turn on precise language. Internalize these:

• Some = at least one. 'Some A are B' also means 'Some B are A.'
• Most = more than half. 'Most A are B' does NOT reverse to 'Most B are A.'
• All / every / only = universal, no exceptions. 'Only A are B' means 'If B, then A' (a conditional, easy to flip incorrectly).
• If two 'most' statements overlap ('Most A are B' and 'Most A are C'), you can infer that at least some B are C — because more than half of A being each guarantees overlap.

When a stimulus stacks these, valid inferences are usually about overlaps, not sweeping generalizations. Trap answers love to upgrade a 'most' into an 'all' or reverse a conditional. Now practice.

Eight Original Inference Questions

Work each one untimed first. Read the stimulus, commit to an answer and — this is the key habit — write down why the other four are wrong before you check. The explanations below model exactly that process. (Full stems and choices are in the practice section of this page.)

How to Review Inference Questions (This Is Where Improvement Happens)

Getting a question right isn't the goal; understanding why the four wrong answers are wrong is. For every inference question you miss — or get right for the wrong reason — do this:

1. Categorize your miss. Was the trap too-strong, out-of-scope, reversed, half-right, or plausible-but-unproven? Keep a tally. Within a week you'll see your personal trap pattern.
2. Re-derive the credited answer from the text. Point to the exact sentence(s) that guarantee it. If you can't, you found the answer by elimination — fine, but note it.
3. Prove each wrong answer wrong in one phrase. 'Reverses the conditional.' 'Adds a cause.' 'Upgrades most to all.' If you can't articulate it, you don't own the question yet.

Then move to volume on the real thing. Once the method is automatic, drill official LSAC questions on LawHub under timed conditions — that's the only way to calibrate to actual test difficulty and pacing.