Last updated July 07, 2026 · Reviewed by the AnvayaPrep team
Introduction
Inference Questions ask you to identify what must be true, what is most strongly supported, or what can be properly drawn from the information in the stimulus. Unlike argument-based question types that require evaluating or manipulating a conclusion, inference questions treat the stimulus as a set of facts and ask what those facts logically establish.
This unit covers the full range of inference tasks on the LSAT: Must Be True questions, Most Strongly Supported questions, soft inference variations, inference from conditional chains, inference from quantified statements (all, some, most, none), inference from causal claims, inference from comparisons, inference from statistics and rates, inference with analogies, inference in two-speaker dialogues, and the critical skill of avoiding outside assumptions. The unit also addresses the common wrong answer patterns that trap test-takers: extreme answers, new information traps, weakly supported conclusions, and answers that reverse or go beyond the stimulus facts.
Inference questions account for approximately 15 to 20 percent of Logical Reasoning questions on a typical LSAT. Their defining characteristic is conservatism: the correct answer is almost always a modest, carefully hedged statement that does little more than restate or combine facts already given. Students who expect dramatic or insightful inferences consistently choose wrong answers.
Learning Objectives
- Identify the logical standard required by different inference stem variations: "must be true," "properly inferred," "follows logically," "most strongly supported," and "could be true."
- Apply the principle that a valid inference adds no information beyond what is logically contained in the premises, making no extrapolations or value judgments.
- Combine multiple statements from the stimulus to derive compound inferences that are individually supported by each contributing piece of evidence.
- Draw valid inferences from conditional chains by following arrows and applying the contrapositive without committing illegal reversals or denials.
- Infer from quantified statements (all, some, most, none) using the formal rules governing each quantifier's logical implications.
- Recognize and avoid the most common wrong answer patterns: extreme language, outside knowledge introduction, answers that reverse stimulus relationships, and answers that require additional assumptions.
- Distinguish between what must be true, what is most strongly supported (probable), and what could be true (merely consistent with the stimulus).
High-Yield Concepts
| Inference Type | What It Requires | Common Pitfall |
|---|---|---|
| Must Be True | Conclusion is guaranteed; no possible scenario where stimulus is true but answer is false | Choosing an answer that is likely true but not guaranteed |
| Most Strongly Supported | Best-supported answer; allows for probability, not certainty | Choosing an extreme statement or one requiring outside knowledge |
| Inference from conditionals | Follow the arrow; apply contrapositive; never reverse or deny without contrapositive | Illegal reversal (If A then B does not mean if B then A) |
| Inference from quantifiers | All + All = All; Some + All = Some; Most + Most does not guarantee overlap | Overgeneralizing from "most" to "all" or from "some" to "most" |
| Inference from causal claims | Correlation established does not guarantee causation in the inference; stick to what is stated | Treating correlation as established causation |
| Fact-set inference | Combine overlapping claims; look for guaranteed implications of combined information | Introducing information not in the stimulus |
| Soft inference (could be true) | Any answer not contradicted by the stimulus | Forgetting this standard and looking for something guaranteed |
Approach inference questions by asking: "What do these facts guarantee?" rather than "What seems reasonable?" The correct answer will feel almost obvious or underwhelming. If an answer choice feels surprising or insightful, it probably requires an assumption not stated in the stimulus.
Valid inference checklist: (1) Is every piece of information in the answer supported by the stimulus? (2) Does the answer go beyond the stimulus in any way? (3) Could the stimulus be true while the answer is false? If you answered "no, no, no" to these three questions, the answer is likely correct.
Study Strategy
Begin this unit by mastering the distinction between "must be true" and "most strongly supported" standards. Practice identifying which version a stem requires before reading the stimulus or answer choices.
Study conditional inference and quantifier inference as formal rule systems rather than intuition-based tasks. The rules for all, some, most, and none follow predictable patterns. Drill these patterns until they are automatic: All A are B + All B are C = All A are C; Some A are B + All B are C = Some A are C; Some A are B does not combine with Some B are C to yield anything about A and C.
Work through inference-with-analogies, inference-with-cause-and-effect, and inference-from-statistics topics carefully, as these involve applying domain-specific reasoning patterns within the formal "what must be true" standard.
Practice the pre-phrasing habit: before reading answer choices, predict what the correct answer should say. Even an approximate prediction eliminates three to four wrong answers immediately.
Think of the correct answer to a Must Be True question as a shadow: it is cast by the stimulus facts and cannot exist without them, but it also cannot be larger than the light source. The answer cannot go beyond what the stimulus contains.
Common Mistakes
Selecting answers with extreme language. Words like "always," "never," "all," "none," "impossible," and "certainly" signal a claim stronger than what most stimuli can guarantee. When the stimulus uses hedged language, the correct inference must also be hedged.
Introducing outside knowledge. Inference questions prohibit reliance on information not in the stimulus. If you find yourself thinking "but I know that in real life..." you are about to choose a wrong answer. Everything needed to justify the correct answer is in the stimulus.
Reversing conditional relationships. From "If A then B," you can infer "If not B then not A" (contrapositive), but not "If B then A" (illegal reversal) or "If not A then not B" (illegal denial). Many wrong answers exploit illegal reversals.
Combining "some" statements incorrectly. "Some A are B" and "Some B are C" do not jointly imply "Some A are C." The "some" overlap might not include the same B's. This is one of the most commonly missed quantifier combinations.
Treating "most strongly supported" as "must be true." The most-strongly-supported standard allows for a probabilistic best answer, not a logically guaranteed one. On harder questions, this distinction determines whether you look for a certain inference or a well-supported one.
Exam Tips
After reading the stimulus for an inference question, try to state in one sentence what the passage most clearly establishes. This pre-phrasing narrows the answer choices before you read them.
Wrong answers in inference questions typically fall into five categories: (1) outside information added, (2) extreme language unjustified by the stimulus, (3) reversal of a stated relationship, (4) negation of a stated relationship, or (5) an answer that is merely consistent with the stimulus rather than supported by it. Use these categories to eliminate wrong answers systematically.
For two-speaker inference questions, treat both speakers' statements as true facts and combine them. Do not infer based on what one speaker believes; draw conclusions from the combined set of facts.
Do not select an answer simply because it sounds reasonable or because you believe it to be true in the real world. The correct answer must be supported exclusively by the stimulus. An answer that introduces a new concept, adds a new claim, or requires any knowledge beyond what is explicitly stated is wrong, no matter how plausible it sounds.
Sign up free to keep reading
Create a free AnvayaPrep account to finish this LSAT guide on Inference Questions — plus flashcards and practice questions.