The 14 LSAT Logical Reasoning Question Types, Ranked by Frequency

Logical Reasoning is now the center of gravity of the LSAT. Since the August 2024 administration, the scored multiple-choice test contains two LR sections, one Reading Comprehension section, and one unscored variable section — which means roughly two of every three scored questions are Logical Reasoning. If you are not sure how to spend your study hours, that arithmetic answers the question before any other consideration does.

The bad news is that LR punishes aimless practice harder than any other section. Ten hours of mixed drills with no type awareness will move your score by almost nothing, because you are relearning the same three or four errors while paying tuition to the remaining ten question types. The good news is that LR is also the most learnable section on the exam. The question types are finite, each has a characteristic stem, and each has a decision procedure you can execute under pressure.

What is actually on the LSAT right now

Before the type-by-type breakdown, the format facts that determine how much LR matters to your score. These are the numbers to hold in your head, and they are the numbers most prep material still gets wrong.

Current U.S./Canada multiple-choice format. A 10-minute intermission follows section 2; LR Argumentative Writing is a separate, unscored 50-minute task taken online.

One more change worth knowing if you are registering now: beginning with the August 2026 administration, LSAC requires in-person testing at a Prometric center for almost all test-takers, with narrow exceptions for approved medical accommodations or extreme travel distance (documented as more than 180 miles or three hours of travel). The questions, the timing, and the 120–180 scale are unchanged — what changed is the room you sit in. That makes full-length practice under timed, distraction-free conditions more valuable than it was, not less.

The question types, ranked by how often they appear

LSAC does not publish an official percentage-by-type table, and independent trackers disagree on exact decimals because they classify borderline questions differently. The ranking, however, is stable across every serious analysis. Treat the percentages below as planning guidance, not law.

The five you must own

1. Necessary Assumption — the negation test is not optional

A necessary assumption is a statement that, if false, collapses the argument. The procedure that beats intuition almost every time is mechanical: take each answer choice, negate it, and ask whether the argument still works. The choice whose negation breaks the argument is the answer. If you have never trained this reflex, your accuracy on these questions is roughly random regardless of how well you understand the stimulus.

The two traps are (a) finding a statement that merely helps the argument, when the question asks what it requires, and (b) choosing an answer that is too strong. An assumption that is not strictly necessary is wrong even when it feels obviously true — necessary, not helpful.

2. Flaw in Reasoning — name the error before you read the options

Flaw questions ask what is wrong with the argument, and the answer choices describe errors in abstract terms. If you read the options first you are pattern-matching on vocabulary. If you identify the error first, the correct answer is almost always a restatement of it in different words.

The errors that carry the most points: reversing a conditional (treating a sufficient condition as necessary), equivocation (a term shifting meaning mid-argument), causal fallacy (correlation treated as cause, reversed causation, or an unconsidered third cause), faulty comparison (comparing groups that are not comparable on the relevant dimension), hasty generalization, and exclusivity errors (assuming that because something has not been proven, it is false).

3. Inference / Must Be True — be conservative

Treat everything in the stimulus as true and select only what must follow. The classic error is picking the answer that is highly plausible, well-supported by the real world, or consistent with the author's view but not strictly required by the passage. Anything that adds a new fact, narrows a set, or smuggles in outside knowledge is wrong.

Practical filter: if a choice would still be true if one of the stimulus's premises were removed, it is not required by the stimulus. "Must Be True" questions are scored on logic, not plausibility.

4. Weaken and 5. Strengthen — attack the gap, not the conclusion

Both are gap questions. Find the assumption connecting the premises to the conclusion, then decide whether your answer weakens or strengthens that link. Weakening the conclusion's wording while leaving the reasoning intact is a common non-answer, as is supporting a claim the argument never needed in order to reach its own conclusion.

There is also a small family of questions — Sufficient Assumption, Evaluate — where the logic runs the other way. A sufficient assumption, if true, guarantees the conclusion: the answer plus the stimulus must leave no alternative explanation standing. For Evaluate questions, your job is to find the choice that is decisive either way — one that, if true, weakens the argument, and one that is false strengthens it.

The nine that show up less, and how to handle them

You still need a procedure for each, but you can spend far less time drilling them.

1. Resolve the Paradox

   Two facts appear to conflict. The correct choice explains both simultaneously rather than picking a side. Test each option by asking whether the paradox survives if it is true. If it does, it is not the resolver.

2. Method of Reasoning

   Describe what the argument does, never what it concludes or whether you agree with it. Wrong answers restate the stimulus's conclusion or its topic; the right answer is about the argumentative technique.

3. Main Point / Conclusion

   The conclusion is what the author is trying to establish — not the last sentence, not the most specific claim. Abstract each candidate into a short noun phrase and pick the one that best fits.

4. Principle (Identify / Apply)

   Identify asks for the general rule that justifies the reasoning. Apply flips it: find the scenario that conforms to the stated principle. In both directions, abstract away from the stimulus's subject matter entirely.

5. Parallel Reasoning and Parallel Flaw

   Translate the stimulus into pure logical form (All A are B; X is A; therefore X is B) before reading the options. Match structure, never topic. Parallel Flaw matches the error rather than the conclusion.

6. Role of a Statement / Point at Issue

   Role questions ask for the function of a specific sentence — premise, conclusion, background, or rebuttal. Point at Issue questions ask exactly where two speakers come into conflict, which is usually narrower than their overall disagreement.

Pacing, review, and the mistake that costs the most points

With 24–26 questions in 35 minutes you have roughly 80 seconds per question, and the distribution is uneven. Some LR arguments solve in 40 seconds; a dense conditional-reasoning stimulus can eat three minutes. If you handle questions in the order they appear, one hard stimulus can cascade — you lose the next three questions to a compressed clock.

The fix is a skip-and-return policy decided in advance. Pre-phrase a rough answer, and if none of the choices come close, flag the question and move. Reserve the last four or five minutes of the section for flagged questions. Decide this before test day, because deciding it mid-section is how the rule gets abandoned.

The second-biggest leak is skipping the review step. After every practice section, redo each flagged or guessed question untimed and before opening the answer key. The gap between your timed score and your blind-review score is the portion of your score that timing is costing you, and it is the only clean diagnostic separating a knowledge problem from a pacing problem. Keep a wrong-answer journal with three columns: question type, the error you made, and the rule you will apply next time. Reviewing that journal once a week is worth more than five more hours of undirected drilling.

Does your target school care?

LR consistency matters more for some applicant pools than others, which is worth knowing before you decide how many hours LR deserves. Schools with high interquartile ranges — where the 25th-percentile LSAT sits close to the median — are admitting a large share of applicants clustered in a narrow band, so a single unstable LR section can push you below the line. You can see exactly how tight that band is in our LSAT score data by school.

Use the LSAT score percentile chart to see where a target score lands, browse the full school database for admissions figures, and check what a score actually opens up at a given school before you commit study hours to a specific target. If you want to model your chances quickly, the free LSAT score calculator runs against all 196 ABA schools at once, with no signup.

A workable study sequence

1. Weeks 1–2: untimed drills, five question types only. Accuracy over speed. Journal every error by type.
2. Weeks 3–4: add timing. Aim for 2 minutes per question before you compress further — this is where the pacing habits form.
3. Weeks 5–6: the remaining nine types, then mixed sets with no type label so you have to read the stem.
4. Week 7 onward: full-length practice under real conditions. Since August 2026 that means simulating a quiet test center, no phone, no notes.
5. Final week: journal review only. Do not take a new practice test in the last three days; a bad simulated score costs you confidence you cannot rebuild in 72 hours.

That sequence assumes roughly 250–300 total study hours, which is the commonly cited range for a serious three-to-six-month preparation. Adjust the calendar, not the order — the sequence is what produces the score.

Questions we get asked

LSAT® is a registered trademark of the Law School Admission Council, which does not review or endorse this site. This article is educational and is not legal or admissions advice. Score and policy figures reflect LSAC publications available as of October 2026 and change — always confirm against lsat.lsac.org before you register.