How to Avoid Reversing “If A, Then B” on a Reasoning Test

Why doesn’t B alone establish A?

‘If A, then B’ guarantees B when A is true; it does not establish A when B is observed. Mark the given fact and proposed conclusion, then check the rule’s direction before choosing an answer. If the rule points from A to B but the item gives B and asks you to conclude A, leave A undetermined unless another premise supplies the missing link. The Stanford Encyclopedia of Philosophy’s “Necessary and Sufficient Conditions” states: A is sufficient for B, while B is necessary for A. Neither says B is sufficient for A.

Consider this original illustration. Suppose a library rule says: if a book is overdue, the library sends a reminder. A reminder arrives, but the message does not identify the overdue-book condition by itself. The overdue book is one possible explanation, but the reminder could concern a book due soon or be a general notice. The stated rule does not rule those alternatives out. So the reminder does not prove that a book is overdue. This does not show which explanation is likely; it shows the rule can hold while the proposed conclusion is false.

Keep the rule’s direction visible, separate stated facts from tempting inferences, and ask whether the conclusion must follow. OpenStax’s “Equivalent Statements” explains that the converse swaps the condition and result, changing “if A, then B” into “if B, then A.” The converse is a new claim. By contrast, reversing and negating both parts gives the contrapositive, “if not B, then not A,” which the text identifies as logically equivalent to the original conditional.

Put A and B in the order given, then compare that order with the proposed conclusion. If the answer moves from B back to A, look in the prompt for an explicit reverse rule, a two-way definition, or other stated condition that closes the gap. Without one, the careful answer is not “A is false”; it is “A has not been established.” A may still be true. The distinction is between a plausible explanation and a conclusion required by the stated rule. This first-pass check is a way to inspect an item, not evidence that using it improves general intelligence or any overall test score.

Sources: Necessary and Sufficient Conditions; 2.5 Equivalent Statements

How can one counterexample expose a reversed inference?

A quick way to test the reverse is to keep the original rule true while checking whether B could happen without A. If that case fits the stated premises, then “if A, then B” does not establish A just because B is given. “Entails” means that a conclusion must be true if the premises are true. One compatible case with B true and A false shows that the rule and observation do not entail A. Consider an original illustration. Let A mean “a library book is overdue” and B mean “the library sends a reminder.” Suppose the rule says, “If a book is overdue, the library sends a reminder.” Test the reverse: “The library sent a reminder, so the book must be overdue.” A reminder might instead announce a due date, correct an account record, or arrive as a courtesy before the due date. The reminder still occurred, while the book was not overdue. The one-way rule does not say overdue books are the only books that receive reminders. So the observation B does not force A. The point is compatibility, not likelihood. These alternatives do not show how often such reminders occur. They show only that the rule leaves room for B without A. A test asking what must follow is asking for necessity under the stated premises. “A is one possible explanation” is weaker than “A must be true.” Familiarity is not proof. A compact comparison makes the direction visible: | Case | A | B | “If A, then B” | |---|---|---|---| | A occurs and B follows | true | true | true | | A is absent while B occurs | false | true | true | The second row is the counterexample to the converse, “if B, then A”: it allows the original rule to remain true while the reverse conclusion fails. For contrast, the original conditional would fail in a different kind of case: A is true but B is false. That is the only truth-value combination that makes a formal conditional false. OpenStax’s “2.5 Equivalent Statements” presents this falsifying case and distinguishes a conditional from its converse; which shows that swapping the parts is not truth-preserving. The practical check is therefore not “Can I think of a reason A might explain B?” It is “Can B occur in any case allowed by the rule while A is false?” If yes, the reverse is not guaranteed. The Stanford Encyclopedia of Philosophy’s “Necessary and Sufficient Conditions” describes the standard relation: A is sufficient for B, while B is necessary for A. Other routes to B remain possible. If the prompt separately states that B can occur only when A is true, or gives a second rule from B to A, then the added premise closes this gap. Without it, keep the narrower conclusion: B is observed; A may be possible, but the original rule alone does not prove it.

Sources: 2.5 Equivalent Statements; Necessary and Sufficient Conditions

What exactly does the original rule guarantee?

Under the standard formal reading, “if A, then B” guarantees B whenever A is true. It does not say A is the only way B can be true. In the library illustration, let A mean “a book is overdue” and B mean “the library sends a reminder.” The rule says an overdue book is enough for a reminder. If a reminder arrives, the rule alone does not establish that the book is overdue: a courtesy notice could explain it. These alternatives are illustrations, not claims about a library. “Necessary and Sufficient Conditions,” in the Stanford Encyclopedia of Philosophy, describes this relationship: A is sufficient for B, while B is necessary for A. A sufficient condition guarantees the result: if A holds, B must hold. A necessary condition must hold whenever the target condition does. If A is the target, B is necessary because A cannot be true without B. Necessary does not mean B guarantees A. The converse swaps the two parts: “if B, then A.” In the library example, that would say every reminder means a book is overdue. This adds a guarantee; it cannot be obtained by reading the original rule backward. The contrapositive reverses and negates both parts: “if not B, then not A.” Here it says that if no reminder was sent, the book was not overdue. In formal logic, this is equivalent to the original conditional. OpenStax’s “Equivalent Statements” defines the converse as swapped and the contrapositive as swapped and negated, and shows the latter is equivalent to the original. “Only if” locates the necessary condition. Under this convention, “A only if B” means A→B: B is required for A, though B may occur without A. Thus “a book is overdue only if the library sends a reminder” makes the reminder necessary for an overdue book. “A if B” means B→A, because the condition following “if” supplies the trigger. So “a book is overdue if the library sends a reminder” asserts that a reminder is enough to establish overdue status. The phrases use the same terms but point in different directions. For a test item, translate the statement into A→B before judging an inference. If the prompt gives B and asks whether A must follow, you need an explicit B→A rule or a definition that supplies both directions. With only A→B, leave the reverse unproved. The Stanford Encyclopedia notes that everyday English “if” can express different relations; follow the item’s convention or context rather than treating conversational wording as an extra reverse rule.

Sources: Necessary and Sufficient Conditions; 2.5 Equivalent Statements

Which inference forms should I compare?

Compare four patterns by keeping the original rule fixed and changing only the extra premise. Imagine a neutral rule: “If a parcel is marked fragile, then it receives careful handling.” Let A mean “the parcel is marked fragile” and B mean “the parcel receives careful handling.” The example is illustrative; its everyday plausibility does not add premises to the formal check. 1. **Modus ponens:** Premise 1: If A, then B. Premise 2: A. Conclusion: B. This follows: the rule guarantees B whenever A holds. In the example, if the parcel is marked fragile and the stated rule applies, careful handling follows. But the form alone does not tell you whether any parcel is marked fragile. A valid step can therefore produce no new answer when its second premise is missing. 2. **Modus tollens:** Premise 1: If A, then B. Premise 2: Not B. Conclusion: Not A. This also follows under the formal rule: if A were true, B would have to be true, which conflicts with the given not-B. In words, if the parcel did not receive careful handling, it was not marked fragile, assuming the rule and observation are both accepted. 3. **Affirming the consequent:** Premise 1: If A, then B. Premise 2: B. Proposed conclusion: A. This is the tempting converse and does not follow. Careful handling could have another explanation: perhaps the parcel was unusually heavy, or a worker chose caution for another reason. The premise says fragile marking is enough for careful handling, not that it is the only route to it. So B may fit A, but it does not establish A. 4. **Denying the antecedent:** Premise 1: If A, then B. Premise 2: Not A. Proposed conclusion: Not B. This does not follow either. The parcel might still receive careful handling for one of those other reasons. Here the rule’s trigger is absent, but that does not cancel every other possible cause of B. Affirming the consequent reverses the direction from B to A; denying the antecedent starts with not-A and wrongly concludes not-B. Both fail because the original one-way rule leaves room for cases in which B occurs without A. OpenStax’s “2.5 Equivalent Statements” makes the same formal comparison: it identifies the converse as “if B, then A,” while the contrapositive, “if not-B, then not-A,” is logically equivalent to the original conditional. The contrapositive is a rewritten statement, with its parts switched and negated. Modus tollens is an argument pattern: it uses the original rule plus not-B to infer not-A. They connect, but they are not the same label. Underline the premise supplied and check whether the conclusion follows from it and the rule. A plausible conclusion can remain unsupported, while a valid pattern is useless without its required premise. This audits the argument.

Sources: Advanced Mathematical Study and the Development of Conditional Reasoning Skills; The Effect of Perception of Semantic Relations in Conditionals on Conditional Reasoning; 2.5 Equivalent Statements

How should the check adapt to a test’s question type?

Read the task verb before judging the answer. “Must follow” asks whether a conclusion is true in every case consistent with the premises. “Could be true” asks whether at least one such case permits it. “Most likely” asks for a probability judgment or best explanation, which may depend on evidence beyond a formal rule. These are different standards. A counterexample defeats “must follow” by showing one permitted case where it fails; it does not show the conclusion is unlikely. For example, take the rule “if a library book is overdue, the system sends a reminder.” If a question says a reminder was sent and asks what must follow, an overdue book does not qualify: the rule leaves open other reasons for a reminder. If it asks what could be true, the book might be overdue, unless another premise rules that out. If it asks what is most likely, the rule alone does not settle the answer. We would need evidence about other causes. This is an illustration, not evidence about any library system. “Must,” “necessarily,” or “follows from the statements” signals entailment: keep only conclusions guaranteed by the premises. “May,” “could,” or “is consistent with” asks possibility: reject an answer only if it conflicts with the premises. “Probably,” “best explains,” or “most likely” invites a graded judgment. A possible alternative can affect that comparison, but its mere existence does not make the proposed explanation impossible. Read the whole prompt; another instruction may set the standard more clearly. The opened LSAT Grouping Must Be True Questions guide illustrates one task-specific treatment. It distinguishes conclusions that must be true from those that could be true and teaches rule direction, contrapositives, and rule chaining for legacy grouping games. It is not a study showing score improvement, or evidence that its game procedure transfers to every reasoning test. For another format, identify the requested standard, then follow that item’s stated rules and instructions. The philosophy article “Indicative Conditionals” reviews competing accounts of ordinary “if” statements, including approaches connecting them with conditional probability and relevance. Its discussion cautions against treating every conversational conditional as though context were irrelevant. That helps explain why a real-world “probably” question may need background evidence, while a formal “must follow” item is bounded by its premises. Likelihood is not entailment. Name the standard, then ask whether the answer is guaranteed, allowed, or more likely than its alternatives.

Sources: LSAT Grouping Must Be True Questions; Conditionals As Representative Inferences

When does everyday wording change the first-pass reading?

First identify what kind of “if” the question gives you. A definition may make both directions true; a permission rule may state a one-way boundary; a causal statement may describe a tendency; and a second premise may explicitly provide the reverse. Context helps identify the relation but does not silently add a premise. If the item asks what follows from a stated conditional alone, keep the reading one-way. If it asks what is likely, use supplied evidence and state a probability judgment, not a deduction. Imagine a form defines an “eligible applicant” as an adult resident of the district. By stipulation, eligibility holds if and only if both conditions hold: meeting them is sufficient, and eligibility requires them. The two-way link comes from the definition, not “if” alone. Compare a permission rule: “If a visitor has a pass, the visitor may enter.” This permits entry but does not say every admitted visitor must have a pass; staff might admit others. Reversal needs extra wording or a premise ruling out other routes. Causal wording differs. Imagine a prompt says, “If the motion sensor detects movement, the alarm sounds.” Hearing the alarm does not establish that the sensor detected movement: smoke or a manual switch could also trigger it. With facts about how often each cause operates, the alarm may be evidence for a likelihood judgment. It still does not create certainty from the one-way rule. A counterexample shows the conclusion is not guaranteed, not which cause is more probable. Now compare two premises: “If the library is open, the lights are on” and “If the lights are on, the library is open.” Together they state both directions, so the lights being on supports the library being open within these rules. Search the stem for the second link; do not supply it yourself. The Stanford Encyclopedia of Philosophy entry “Necessary and Sufficient Conditions” explains formal conditional inference while discussing how English “if” can express inferential, explanatory, or evidential relations. It supports checking the intended relation rather than treating conversational conditionals as interchangeable. “Conditionals As Representative Inferences” reviews theories considering conditional probability and relevance; it is theoretical, not a settled rule for every sentence. “The Effect of Perception of Semantic Relations in Conditionals on Conditional Reasoning” reports two experiments with 170 college students each. Results differed across contingency, causal, obligation, permission, and definition conditionals; one inference form performed better for permission and definition statements. Its abstract does not validate this checklist or explain performance on every test. These sources support a modest practice: identify whether the task asks necessity or likelihood, and use only reverse links the prompt gives.

Sources: Necessary and Sufficient Conditions; Conditionals As Representative Inferences; The Effect of Perception of Semantic Relations in Conditionals on Conditional Reasoning

Why does a reverse inference feel plausible?

A reverse inference can feel reasonable because B may be informative about A in context. Suppose the rule says, “If a library book is overdue, the library sends a reminder,” and a reminder arrives. An overdue book is one possible explanation; a reminder might also be sent before the due date or after a renewal request. These illustrate that the rule says what follows when A is true, not that A is the only route to B. Three confidence levels help keep the distinction clear. “Possible” means an explanation fits the information given. “Probable” means evidence or background knowledge makes it more likely than alternatives. “Necessary” means it must be true if the stated premises are true. A reminder makes an overdue book possible under the rule; it does not make it necessary. Reliable information that this library sends reminders only for overdue books could support the reverse inference. Without it, do not silently insert that fact. A conversational “if” may invite more than a bare logical reading. Speakers often choose a conditional because they regard its clauses as connected: one may cause the other, make it relevant, or provide evidence for it. “Conditionals As Representative Inferences” discusses accounts in which a conditional’s appropriateness depends not only on how often the consequent follows the antecedent, but also on a relevant connection between them. This is a theoretical proposal about meaning, not evidence that every ordinary “if” expresses a reliable causal rule. The publisher abstract for “The effect of perception of semantic relations in conditionals on conditional reasoning” reports two experiments with college students, each including 170 participants assigned across five conditional types: contingency, causal, obligation, permission, and definition. One measured perceived relations; the other tested four inference forms. The abstract reports differences in perceived necessity across types and higher modus-tollens performance for permission and definition conditionals than for contingency, causal, and obligation conditionals. It also reports that performance rose with perceived necessity of the consequent. The accessible abstract omits detailed procedures and effect sizes, and the experiments did not validate this checking routine. For a test item asking what follows from a premise, use the formal standard requested. For a real-world question asking what probably happened, consider background evidence, but identify it as evidence rather than a hidden premise. “Necessary and Sufficient Conditions,” in the Stanford Encyclopedia of Philosophy, explains the standard formal relation and challenges arising from varied English uses of “if.” Context may make a reverse guess plausible or likely, but an explicit reverse rule or adequate evidence is needed to justify moving from B back to A.

Sources: Conditionals As Representative Inferences; The Effect of Perception of Semantic Relations in Conditionals on Conditional Reasoning; Necessary and Sufficient Conditions

What can a reasoning study say about improvement?

Research supports a narrow conclusion: conditional reasoning can change on particular tasks, and explanations can sometimes help people carry a solution to a related problem. It does not establish that this article’s direction-check routine has been tested, or that using it raises general intelligence. “Transfer of conditional reasoning: effects of explanations and initial problem types” reports three experiments using Wason selection tasks. In the first, explanations helped transfer from abstract problems to later abstract problems, while verbalization instructions did not. Across the experiments, transfer followed explanations for some pairings, including between abstract and thematic problems; it was stronger for problems of the same type and greater from abstract than thematic starting problems. Thus, an explanation supported performance beyond the first item in some conditions. The outcome was transfer among selection-task problems, not broad intelligence. “Advanced Mathematical Study and the Development of Conditional Reasoning Skills” followed students in five UK schools taking post-compulsory mathematics or English literature. Of 124 students recruited at about ages 15 to 18, 82 completed both assessments over an academic year. The mathematics group changed more than the literature group on conditional-inference tasks, despite no explicit teaching of conditional logic. But the shift was toward what the researchers called a defective conditional interpretation: students more often accepted an inference from a condition’s result back to its trigger and rejected some other inferences. Students chose their subjects, so the comparison was not randomized; change did not show a clear move toward the formal interpretation. It cannot establish a general reasoning gain. “The effect of perception of semantic relations in conditionals on conditional reasoning” reports two experiments with college students. Each included 170 participants assigned across five types of conditional statements; one experiment assessed perceived relations and the other four inference forms. The publisher abstract reports higher modus-tollens performance for permission and definition statements than for contingency, causal, and obligation statements, with performance related to perceived necessity. This suggests wording and perceived meaning matter. The abstract lacks effect sizes and detailed procedures, and did not test this checklist. Together, these findings permit a modest practice claim: explanations and task experience may help on some related conditional-reasoning problems, but transfer depends on what is practiced and how a problem is framed. They do not validate this routine or show far transfer to broad ability. Practice the operation on fresh examples to learn it; do not treat success on familiar items or a browser-test retake as evidence that general intelligence or an IQ estimate has increased.

Sources: Transfer of Conditional Reasoning: Effects of Explanations and Initial Problem Types; Advanced Mathematical Study and the Development of Conditional Reasoning Skills; The Effect of Perception of Semantic Relations in Conditionals on Conditional Reasoning

What should I do after a missed item?

After a missed conditional item, write the rule, given fact, and proposed conclusion separately. Check direction. If the rule says ‘if A, then B’ and the item gives B before proposing A, it has moved from result to possible cause. Look for a second premise licensing that move, such as ‘if B, then A,’ or a two-way definition. If neither appears, record that A is not established. This gives you a reason to correct the answer instead of memorizing which choice was wrong.

Consider an example. A notice says, ‘If a library book is overdue, the library sends a reminder.’ Suppose a reminder arrived and the question asks whether the book must be overdue. Parse: overdue book → reminder; given: reminder; proposed: overdue book. The move reverses the arrow. Test an alternative allowed case, such as a reminder sent before the due date. This does not show how libraries usually behave; it shows that the rule leaves another explanation open. So overdue status is unproved. If asked what is likely, you need evidence about library practices, not just this rule.

Use four review moves: parse the premise, classify the inference, test an alternative, and state the narrow conclusion. Note whether the question asks what must follow, could be true, or is likely. An alternative case can defeat necessity while leaving probability unanswered. OpenStax’s ‘Equivalent Statements’ distinguishes a conditional from its converse and shows that the contrapositive, formed by reversing and negating both parts, is equivalent to the original. Reversing alone is not contraposition.

After reading an explanation, create a fresh example. Write its A → B rule, supply B, and decide whether A follows. Then test not-B. In three Wason selection-task experiments, ‘Transfer of Conditional Reasoning: Effects of Explanations and Initial Problem Types’ found explanations supported transfer in some conditions, generally more between problems of the same type; verbalization alone did not consistently produce transfer. Its accessible abstract gives neither sample sizes nor full methods. The experiments did not test this review routine or broad intelligence, so treat a fresh example as practice of this operation, not proof that general ability changed.

Keep a short review record: the rule, inference form, requested standard, and missing link or counterexample. Do not copy protected items or collect answer keys. A useful explanation shows why a conclusion follows, fails to follow, or remains open under the premises. If it depends on an unstated assumption, separate it from the printed rule. Your note might read: ‘B was given; A was proposed; no reverse rule appeared, so A was not required.’ This makes the check reusable and preserves the difference between a plausible explanation and what the premises guarantee.

Sources: Transfer of Conditional Reasoning: Effects of Explanations and Initial Problem Types; 2.5 Equivalent Statements

What is a responsible next step?

Use the direction check to review low-stakes practice, then interpret any result only as performance under that test’s conditions. One missed conditional item does not establish general ability. Record the rule, the fact supplied, and the proposed conclusion; ask whether the conclusion follows, could be true, or is merely likely. Keep those standards separate when reviewing an answer or trying a fresh example. The aim is to make the reasoning step visible, not to turn one response into a judgment about identity, worth, diagnosis, or fixed potential. Test IQ Free offers a free 50-question reasoning test with worked answers and a provisional estimated IQ range. Its conversion is unvalidated and has no validated adult reference population, so the range is neither a normed comparison nor a confidence interval. The optional $9 report explains the same provisional result; payment does not add validity. For low-stakes practice, take the free test at /test and review its worked answers. If an educational, clinical, or employment decision depends on cognitive assessment, seek a qualified professional assessment. Treat this item’s outcome as performance under defined conditions, not as a prediction of an overall score or a measure of who you are.

Questions readers ask