LSAT Logical Reasoning

Logical Reasoning (LR) is the largest component of the current LSAT by section count: two of the three scored multiple-choice sections use LR, and the experimental section may also be LR. Each section lasts 35 minutes. Questions present a short argument, a set of facts, or a disagreement, then ask you to perform a specific reasoning task. You do not need to know law or the subject matter; the answer must follow from the stimulus and the question's instruction.

Read the argument before judging it

Most arguments contain evidence and a conclusion. The conclusion is the point the speaker wants you to accept; premises are the reasons offered for it. Some stimuli also include background, an intermediate conclusion, an opposing view, or a rule. Before looking at choices, state the main conclusion in your own words and note what the premises actually establish.

Consider: “The city replaced its paper parking permits with an app last month. Complaints about lost permits have dropped since then. The app should be used for all city services.” The first sentence provides context, the second reports an outcome, and the third is the main recommendation. The evidence concerns one narrow problem, while the conclusion covers every city service. That scope gap may matter for a strengthen, weaken, or assumption question.

Words such as therefore, thus, so, consequently, and hence often introduce a conclusion. Because, since, and given that often introduce support. These are useful clues, not guarantees. A sentence beginning with “since” can be part of a conclusion, and a conclusion can appear first. Ask which claim the speaker is trying to establish rather than relying on a signal word alone.

The stem defines the task

An LR stimulus can support several different questions. One asks for a necessary assumption, another asks which statement would strengthen the reasoning, and a third asks which flaw is present. The same argument does not have one universal “best answer.” Read the stem closely and translate it into a job: prove, require, support, weaken, explain, infer, parallel, or describe.

A strong routine is to read the stimulus, identify the conclusion and support, read the stem, make a short prediction, then evaluate choices. Reading the stem before the stimulus can sometimes help you focus, but the full argument still needs to be understood. The prediction does not have to be a polished sentence. It may be “need evidence other services share the same permit problem” or “alternative cause for the drop.”

Assumption questions

A necessary-assumption question asks what the argument requires. The conclusion may fail without the answer choice, even though the choice alone does not prove it. Use the negation test: negate a candidate and ask whether the reasoning loses an essential link. Avoid options that merely make the conclusion more plausible or repeat evidence already given.

A sufficient-assumption question asks what would make the conclusion follow. Add the choice to the premises and see whether it closes the gap completely. A sufficient answer can be stronger than what someone would normally assume. The test is whether the argument becomes valid, not whether the added rule sounds modest.

Original example: “Every approved bridge repair in the county uses the new inspection checklist. This bridge repair used the checklist. Therefore, it was approved.” The premise establishes approved → checklist, but the conclusion reverses the direction. A sufficient assumption is checklist → approved. A necessary assumption might be that no unapproved repair can use the checklist, which is the same missing reverse link in this constructed example. Without that link, use of the checklist might occur for other reasons.

Necessary and sufficient assumptions are often confused because both fill gaps. Ask two different questions. For necessity: would the argument break if this claim were false? For sufficiency: if this claim were true, would the conclusion have to follow? One choice can be necessary without being sufficient, and a strong sufficient assumption can also be necessary in the constructed argument, but the task determines the standard.

Strengthen and weaken

A strengthen answer adds support to the argument. It need not prove the conclusion beyond all doubt; it must make the reasoning more convincing. A weaken answer makes the conclusion less convincing, often by identifying an alternative cause, a flawed comparison, a missing group, or evidence that points the other way. Neither task asks whether the conclusion is true in the real world.

Take the parking-app argument. A strengthening choice might say that several other city services lose records through paper forms and that the app's audit feature prevents the same problem. That connects the observed benefit to the broad recommendation. A weakening choice might say that complaints fell because the city stopped issuing temporary permits during the month, so the app did not cause the decrease. These facts affect the reasoning, not merely the topic.

For causal arguments, ask what else could produce the outcome, whether the cause preceded it, whether the sample is representative, and whether the effect could have caused the alleged cause. For a comparison, ask whether the cases differ in a relevant way. For a prediction, ask whether the conditions will continue. The answer need not use technical labels; it needs to change how much the evidence supports the conclusion.

Flaw questions

A flaw question asks you to describe an error in the reasoning. Frequent flaws include treating correlation as causation, assuming that what is true of one member is true of a whole group, generalizing from an unrepresentative sample, confusing a necessary condition with a sufficient one, and rejecting a claim because its alternative has not been proven.

Original example: “Every engineer on the design team can use the modeling software. Therefore, everyone who can use the software is an engineer on the team.” The premise gives team engineer → can use software. The conclusion reverses the conditional. Other employees may also use the software. The flaw description should capture the reversal, not simply say that the conclusion is too broad.

A useful check is to distinguish an invalid inference from a questionable premise. If the question asks about reasoning, an answer that says the conclusion may be false is too general. State how the evidence fails to establish it. If the argument moves from a small survey to all residents, the problem is representativeness; if it treats “some A are B” as “all A are B,” the problem is quantity.

Inference and must-be-true questions

An inference question asks for a conclusion supported by the given information. Choose the strongest statement that must be true or is most strongly supported, according to the exact wording. Do not strengthen the facts. If a stimulus says some library branches open late and every branch that opens late has a security guard, you can infer that some branches have guards. You cannot infer that all branches open late or that every guarded branch is open late.

Quantifiers control the strength of an inference. “All” supports a universal statement in one direction; “some” establishes existence; “most” does not mean all; and “not all” means at least one exception. Preserve those distinctions. Conditional claims can also be combined: if A implies B and B implies C, then A implies C. The reverse conclusion does not follow unless another premise supplies it.

Explain or resolve a discrepancy

Some questions give two facts that appear inconsistent and ask what would explain them. The answer should allow both observations to be true. For example, a store's revenue could rise while its number of customers falls if average spending per customer increases enough. A choice that simply repeats one observation does not resolve the tension.

Separate the apparent contradiction into its components. If a report says a new bus route has many riders but little effect on total car traffic, perhaps most riders previously used another bus line. If a conservation effort protects a forest but the measured bird population falls, a new survey method may have detected previously uncounted habitat differences. The correct explanation reconciles both sides without inventing unsupported details.

Parallel reasoning and structure

Parallel reasoning questions ask you to identify an argument with a similar logical structure. Ignore matching subject matter and compare the form: premises, conclusion, conditional direction, quantifiers, and any flaw. A bird argument and a manufacturing argument can be parallel even when they share no vocabulary. If the original makes a converse error, the parallel answer should make the same kind of move.

Parallel flaw questions add a second requirement: match both the structure and the reasoning defect. Translate the original into a simple template if needed. For example, “All certified auditors receive training; Mina received training; therefore Mina is a certified auditor” has the form A → B; B; therefore A. A matching choice must repeat the invalid reversal, not merely mention certification or training.

A worked strengthen question

Argument: “After a hospital introduced text reminders, missed appointments fell by 18 percent. The hospital should use text reminders for every department.” The evidence is a temporal change at one hospital, while the recommendation generalizes across departments. The gap includes causation and whether departments have similar appointment patterns.

Choice A: “Some patients prefer phone calls.” This may matter for implementation but does not directly establish whether reminders reduce missed appointments. Choice B: “No other scheduling policy changed during the trial, and departments with different appointment types experienced similar decreases.” This supports causation and makes the result more generalizable, so it strengthens the conclusion. Choice C: “The hospital has used text reminders for five years.” Duration alone does not show effectiveness. Choice D: “The text vendor also sells billing software.” This fact is unrelated to the reasoning.

Choice B is not necessarily proof that reminders will work in every imaginable department; it provides support, which is the task. The distinction between “strengthens” and “proves” prevents rejecting a useful answer because it is not conclusive. It also explains why an answer can be relevant but still weaker than the best option.

A worked inference question

Facts: “All projects approved by the safety board include a site inspection. Some projects with site inspections also receive a second review. No project lacking a documented risk plan can be approved.” What must be true? Approved projects have site inspections. Approved projects also have documented risk plans. You cannot infer that every project with an inspection is approved, because the first statement goes only from approval to inspection. You also cannot infer that all approved projects receive a second review, since only some inspected projects do.

This example illustrates a useful discipline: mark the direction of each rule, combine only compatible claims, and retain qualifiers such as some or all. A choice that sounds likely based on ordinary practice is not enough. The question asks what follows from the statements as given.

Why wrong answers are tempting

A wrong answer can feel attractive because it uses the same subject words as the argument. Another may be true but answer a different question. A third may strengthen the conclusion when the stem asks for a necessary assumption. A fourth may introduce an extreme claim that the evidence does not support. When two answers seem close, restate the task and describe what the argument needs before rereading the options.

Do not bring outside assumptions into the stimulus. If an argument discusses a public survey, you may know that surveys can be biased, but an answer must identify a plausible bias tied to the facts or reasoning. If it discusses a scientific study, you need not know the science. Focus on what was measured, whom it included, when it occurred, and what conclusion the author draws.

Timing without careless reading

The 35-minute section clock requires flexible pacing. Some questions can be answered quickly, while parallel reasoning or dense stimuli may take longer. If a question is consuming time without a clear path, eliminate what you can, mark it, and continue. Return if time remains. Answer every question because a wrong answer is not penalized.

Do not skim the stimulus so aggressively that you miss a qualifier or conclusion. A fast initial read followed by several rereads often costs more time than one careful read. Use a short note such as “causal claim, one before-after example” or “principle then exception.” Keep the note brief and use the answer choices to test a clear prediction.

After a timed section, review it untimed. If an item was correct but guessed, identify the uncertainty. If you missed an item, solve it again before reading an explanation. Then compare your reasoning with the correct answer and record the mistake. This untimed review teaches the skill; timed attempts reveal whether you can apply it under pressure.

A repeatable LR checklist

1. [object Object]
2. [object Object]
3. [object Object]
4. [object Object]
5. [object Object]
6. [object Object]

Logical Reasoning improves when the process becomes consistent and the review is specific. Memorizing a list of flaw names can help you describe an error, but it is not a substitute for reading the argument. Use the name only after you see the structure. On each new question, let the stimulus and stem determine the task.

Sources

• [object Object]
• [object Object]
• [object Object]