Inferences
The most-missed type on the section. The bar is logical necessity, not plausibility.
~28 min · prequestion, worked examples, retrieval practice
An inference question never asks what probably follows — it asks what cannot fail to follow, and those two standards select different answers. Three of the four choices will be things a sensible person might believe after reading the passage; exactly one is a thing the passage forces on you. Read for plausibility and you arrive at a coin flip between two good-sounding options. Read for necessity and you arrive at one.
Before you read on
Two or three questions on exactly what this lesson teaches. Being wrong here is fine — it's the fastest way to find out what to pay attention to next.
In a recent overwintering trial, every hive that was fed the experimental supplement survived the winter. Hive 12, one of the hives in the trial, did not survive the winter.
Before any teaching: which statement must be true if both sentences are true?
A passage reports a research finding and then stops mid-sentence: "However, ______". Two choices are left standing, and both are consistent with the passage: (i) a sweeping claim that the finding overturns the field's long-standing consensus, and (ii) a narrow claim that the finding does not by itself identify the mechanism involved. Which is more likely to be the keyed answer, and on what grounds?
Between 2005 and 2020, the proportion of the city's commuters who cycled to work rose from 4 percent to 11 percent. Over the same period, the city's population fell by roughly a third.
Which statement is guaranteed by the text?
Foundations — what an inference actually is
Skip this block if you can already state, without hesitating, the difference between a conclusion that follows from a text and a conclusion that merely fits it. Everything after this depends on that distinction, so it is set out here in full for anyone who has never been taught it.
An inference is a conclusion the facts force on you. Not the most likely explanation, not what a reasonable person would assume, not what is probably going on — the thing that has to be true if every sentence you were given is true. If you are told that every book on the shelf is blue, and that the dictionary is on the shelf, you know the dictionary is blue. Nobody told you that; you know it anyway. What you do not know is how many books there are, who put them there, or whether blue books are unusual. Those are all reasonable things to wonder about, and none of them follow.
On the Digital SAT this appears as a short passage — usually forty to a hundred and twenty words — whose final sentence stops in the middle, followed by the prompt "Which choice most logically completes the text?" Each of the four choices is a phrase that finishes that sentence. The passage almost always has the same shape: it sets something up, then reports a finding or a complication, and the blank sits immediately after the complication.
The sentence containing the blank nearly always opens with a connective: "Therefore," "This suggests that," "This finding indicates that," "However," "In other words," "For this reason." That word is not decoration. It states the logical job the missing phrase has to do — draw a consequence, mark a limit, restate. A reader who takes in the passage and ignores the connective is guessing at what the completion is even supposed to accomplish.
One question decides every choice: could this be false while every sentence in the passage stays true? If yes, it is out, no matter how sensible it sounds. If no, it is the answer. That test is the whole skill. The rest of this lesson is about running it quickly, and about not flinching when the guaranteed answer turns out to be the dullest of the four.
What the question is actually testing
Information and Ideas is about 26 percent of the Reading & Writing section — that weighting is College Board's own, from the published assessment framework. Inferences is one of its four skill points. Prep-industry score reports consistently place it among the most-missed R&W types; College Board publishes domain weightings but not per-skill miss rates, so treat "most-missed" as an industry claim rather than an official statistic. Either way, no other R&W type rewards a fixed procedure this heavily.
The standard is not "which of these is most reasonable" but "which of these is guaranteed." Be clear about the provenance of that framing: College Board describes the task as drawing a reasonable inference and most logically completing the text. The word "guaranteed" is Meridian's operating rule, derived from how the keyed answers actually behave across released material — not a phrase College Board uses. It is the rule that makes the keys predictable, and the mechanism block below explains why the test has no real alternative to it.
The first consequence is counterintuitive: the correct answer is very often the least impressive of the four. It hedges — "may," "at least some," "does not by itself establish," "in this sample." The wrong answers frequently sound more informative, because they are claiming more, and claiming more is exactly what makes them unguaranteed. A strong reader who has been rewarded all their life for extracting the biggest available conclusion is being trained in the wrong direction here.
The second consequence is that outside knowledge is a liability on this question type in a way it is nowhere else on the test. A student who already knows the biology, the economics, or the history behind a passage will quietly fill the gaps the passage left open and then pick the choice that matches what they know rather than what they were given. If you find yourself justifying a choice with a fact that is not on the screen, that is the trap closing.
The method, in full
Step 1 — read the whole passage, then go back and read the connective in front of the blank on its own. Name the job: CONSEQUENCE ("therefore," "this suggests that," "this finding indicates that") means the completion states what follows from what was just reported. CONTRAST ("however," "but," "yet") means it states a limit on, or a complication for, the idea already given. RESTATEMENT ("in other words," "that is") means it says the same thing in plainer terms and adds nothing. CAUSE ("because," "since") means it supplies the reason for what precedes.
Step 2 — strip the passage to two or three bare claims, adjectives removed. Typically: what was assumed or standardly believed, what was found, and what complicates the finding. Do this in your head in one pass. It matters because the hardest items are entailed by two of those claims combined, never by one, and the distractors are built out of single claims.
Step 3 — predict the completion in your own words before you look at the choices. It does not have to be elegant. "So the light isn't about predators" is a perfectly serviceable prediction. This is the same discipline Words in Context runs on, and for the same reason: predicting first means the distractors have to beat a commitment you have already made, rather than being your first exposure to any candidate answer.
Step 4 — run the necessity test on all four choices, one at a time, in the same words each time: could this be false while every sentence in the passage stays true? Do not stop at the first choice that sounds right. On this type specifically, the plausible-but-unsupported distractor is usually positioned ahead of the key, and stopping early is how it earns its place.
Step 5 — if two choices survive, take the weaker one, provided it still does the connective's job. A claim that says less has fewer ways to be false and so is easier for the passage to force; a claim that says more has to earn every extra word from something actually written. If the weaker survivor does not do the connective's job — it draws no consequence where "therefore" demanded one — then it was never really a survivor.
Mechanism
Why the bar is necessity and not plausibility
A keyed answer has to survive every student who challenges it, every teacher who writes in, and the item's own statistics. Under a plausibility standard, two or three of four choices are defensible at once — and an item with two defensible answers is a broken item. It shows a low or negative point-biserial correlation, because the strongest readers split between the two good choices instead of converging, and items that behave that way get pulled from the pool. The only standard that leaves exactly one defensible choice is entailment: the answer the passage forces rather than the answer it invites. So when the key is the modest claim rather than the interesting one, the test is not being pedantic; it is doing the only thing that produces a scorable question. The same pressure explains the hedging. A narrow claim is cheap to guarantee, so item writers reach for it, because a narrow claim is what survives review. Reading a hedge as "too weak to be the answer" inverts the actual signal.
Worked examples
Fully worked — a consequence completion
- 01Passage: "Biologists have long assumed that the luminous secretion released by the deep-sea shrimp Acanthephyra serves to startle approaching predators. In a recent series of trials, however, shrimp released the secretion just as often when isolated in a tank with no predator present as when a predator was introduced, and in both conditions they released it only after being physically jostled. This suggests that the secretion is ______"
- 02Step 1 — the connective. "This suggests that" is a CONSEQUENCE marker: the completion has to state what follows from the trials just described, not introduce a new topic and not restate the old assumption.
- 03Step 2 — bare claims. (a) The standard assumption: the secretion startles predators. (b) Finding: release rate was identical with and without a predator present. (c) Finding: in both conditions, release followed physical jostling and nothing else.
- 04Step 3 — predict. Whatever sets the secretion off, it is not the predator; it is being knocked about. Something like: "triggered by contact, not by detecting a predator."
- 05Step 4 — necessity test on the tempting overreach: "...of no defensive value at all." Could that be false while (a), (b), and (c) stay true? Easily — the trials show what triggers the secretion, not what it does once released. It could still deter a predator perfectly well. Out.
- 06Step 4 (continued) — necessity test on the prediction: "...triggered by physical disturbance rather than by the detection of a predator." Could that be false while (b) and (c) stay true? No. If detection triggered it, the rates would differ; they did not. If something other than jostling triggered it, releases would have occurred without jostling; they did not.
- 07Step 5 — check the hedge level. The surviving claim is only about the trigger, and says nothing about function. That narrowness is exactly why it survives. Answer: triggered by physical disturbance rather than by the detection of a predator.
One step hidden — a contrast completion
- 01Passage: "The usual explanation for the collapse of the region's oyster beds is over-harvesting, and sediment cores from the estuary support part of that account: oyster abundance began falling in the 1870s, the decade commercial dredging began. The same cores, however, show that abundance in the neighbouring inlet, where dredging was prohibited throughout the period, fell by a comparable margin over the same decades, which suggests that ______"
- 02Step 1 — the connective. "However" followed by "which suggests that" means the completion has to state what the contrast implies. Restating the finding is not enough; the completion must draw the limit the contrast establishes.
- 03Step 2 — bare claims. (a) Standard explanation: over-harvesting. (b) In the dredged estuary, the decline starts with dredging. (c) In an inlet where dredging never occurred, a comparable decline happened over the same decades.
- 04Step 3 — predict. Something other than dredging drove the inlet's decline, so dredging cannot be the whole story.
- 05Step 4 — necessity test on the overreach: "...dredging had no effect on oyster populations." Could that be false while (b) and (c) stay true? Yes — dredging could have contributed in the estuary while something else drove the inlet. Out for overreach.
Two steps hidden — evidence that cannot discriminate
- 01Passage: "It has been argued that the copper artefacts recovered at the Bandiagara site were imported rather than produced locally, on the grounds that no smelting furnace has been found in the excavated area. Recent isotope work complicates the argument: the copper's trace-element signature matches ore deposits within forty kilometres of the site, but the same signature also occurs in ore from two more distant sources that were traded widely in the period. The analysis therefore ______"
- 02Step 1 — the connective. "Therefore" is a consequence marker, and the subject is "the analysis," so the completion must state what the isotope work establishes — not what the artefacts are, and not what the excavation shows.
- 03Step 2 — bare claims. (a) The import argument rests on no furnace having been excavated. (b) The signature is consistent with nearby ore. (c) The identical signature also occurs in two widely traded distant ores.
- 04Step 3 — predict. Consistent with local ore, consistent with imported ore, so it cannot tell the two apart.
Solve alone
- 01Passage: "Economists studying a 2016 municipal ordinance that raised the local minimum wage compared restaurant employment in the city with restaurant employment in three neighbouring towns that left their wage floors unchanged. Over the following two years, restaurant employment in the city grew by 2 percent, while employment in the comparison towns grew by 6 percent. The comparison therefore suggests that ______" Work through steps 1 to 5 before revealing anything below.
In your own words
In one sentence: why does asking "could this choice be false while every sentence in the passage stays true?" eliminate an answer that a "which of these is most likely?" reading would keep?
Named traps
- The plausible outsider
- The choice is probably true in the world, or true given what you happen to know, but the passage never gives it. This is the most common wrong answer on the type, and the tell is that justifying it requires a sentence you supply yourself. If your defence of a choice begins "well, obviously," the choice is out.
- Quantifier inflation
- The passage's claim with an absolute bolted onto it — all, none, only, never, always, any, proves, cannot. The passage supports "in this trial"; the choice says "in general." Read the quantifiers in every choice before you read the verbs; on a close item, the quantifier is usually where the decision is.
- Half-passage completion
- The choice is genuinely entailed by the setup but ignores the complication the blank sits after — it would have been the right answer if the passage had ended two sentences earlier. Produced almost entirely by not reading the connective in front of the blank and so not noticing that the completion's job is to limit the earlier idea, not extend it.
- Reversed dependency
- The choice swaps what depends on what, or flips the direction of a comparison: the passage establishes that A varies with B and the choice says B is caused by A; the passage says one quantity grew more slowly and the choice says it fell. The content is all from the passage, which is what makes it convincing at a glance.
- Wrong quantity
- Correct reasoning about a real quantity that is not the one the passage measured — a share answered with a count, a rate answered with a total, a change within one group answered as a difference between two groups. The arithmetic in these choices frequently checks out, which is precisely why they are dangerous.
- Recommendation creep
- The passage reports what is; the choice says what someone should do, will do, or must now believe. Prescriptions and predictions are almost never entailed by a description, and a completion that tells researchers what to investigate next is doing something the passage did not license.
The 800-level margin
If you are already reading for necessity and still dropping one or two of these per test, the remaining errors are almost never comprehension. They are four hard variants and three execution slips. This block is that list; nothing in it is a new concept, which is the point.
Hard variant one — two hedged survivors. On the hardest items, both remaining choices are appropriately cautious, so caution stops being a tiebreaker. Decide on scope instead: which one is about the quantity the passage actually measured, and which one covers every alternative the passage raised. If a passage names two confounds and a choice neutralises only one of them, that choice is not a weaker version of the right answer — it is half an answer, and half an answer is unguaranteed.
Hard variant two — the two-hop entailment. Easier items are entailed by a single sentence. The hardest are entailed only by two sentences combined, and the choice that restates one sentence exactly is placed on purpose. Practical rule: if you can verify a choice by looking at one line, be suspicious rather than relieved.
Hard variant three — claim versus establishment. "Researchers proposed," "one hypothesis holds," "the author argues," "if this reasoning is sound." When a passage reports a claim instead of asserting it in its own voice, the completion can only be about that claim, its consequences, or its status — never about the world. Two distractors recur on these: the one that quietly upgrades the hypothesis to a fact, and the circular one, which says the evidence proves the very hypothesis you were told to assume.
Hard variant four — absence of evidence. "The study found no association" licenses "the study did not detect an association." It does not license "there is no association," and it never licenses "the study showed the two are unrelated." The same holds for artefacts and archives: "no furnace has been excavated" is a fact about what has been found, not a fact about what exists.
Execution slips, in the order they cost points. First, direction: a quantity that grew more slowly has not fallen, a share that rose is not a count that rose, and a decline that slowed is not a reversal. Say the direction out loud before you look at the choices. Second, answering the right question about the wrong noun — the passage compares two time periods and the choice compares two groups, or the passage is about what the writer recorded and the choice is about what the reader knew. Third, committing to the first choice that sounds right, which on this type usually means committing to the plausible outsider and never running the test on the choice that would have won.
Time. Each Reading & Writing module gives 27 questions in 32 minutes — about seventy seconds per question, which is arithmetic on College Board's published module structure rather than a recommendation from anyone. The recommendation is Meridian's: settle rule-governed conventions items in well under a minute and let inference items run past ninety seconds, because this is where the extra time converts. The overspend that wastes it is reading the passage a third time; the underspend that costs points is skipping the necessity test on the runner-up.
Retrieval — with feedback on every choice
Ferns reproduce by releasing spores, which must land somewhere moist in order to develop. Because spores are extremely light, wind can carry them hundreds of kilometres, and botanists have documented fern species establishing themselves on newly formed volcanic islands within a few decades of the islands' emergence. Seed plants, whose seeds are heavier and generally travel shorter distances from the parent plant, typically arrive later. Therefore, on a newly formed volcanic island, ______
Which choice most logically completes the text?
INFERENCES — reference card
The standard: could this be false while every sentence in the passage stays true? If yes, it is out.
Read the connective in front of the blank first. therefore/this suggests = consequence. however/but = limit. in other words = restatement. because = cause.
Strip the passage to 2-3 bare claims: what was assumed, what was found, what complicates it. Hard keys need two combined.
Predict in your own words, then test all four choices. Do not stop at the first that sounds right.
Two survivors? Take the weaker one, if it still does the connective's job. Hedged claims are cheap to guarantee.
Read every quantifier before every verb: all / none / only / never / always / proves is where close items are decided.
Outside knowledge is a liability. If the defence of a choice needs a fact not on the screen, it is not the answer.
"Found no effect" licenses "did not detect it", never "there is none".
A reported claim ("researchers proposed", "if this is sound") licenses conclusions about the claim, not about the world.
Direction check: rate is not total, share is not count, slower growth is not decline.Every item on this page is Meridian-original, written to match the Digital SAT's format and difficulty — it is not a real SAT question. The only source that matches the live test exactly is College Board's own Bluebook and Question Bank.
Ferns reproduce by releasing spores, which must land somewhere moist in order to develop. Because spores are extremely light, wind can carry them hundreds of kilometres, and botanists have documented fern species establishing themselves on newly formed volcanic islands within a few decades of the islands' emergence. Seed plants, whose seeds are heavier and generally travel shorter distances from the parent plant, typically arrive later. Therefore, on a newly formed volcanic island, ______
Which choice most logically completes the text?
- ferns are likely to appear before most seed plants do.
Correct. Run the test: could this be false while every sentence stays true? No. The passage gives fern establishment within decades, seed dispersal over shorter distances, and seed plants typically arriving later — the completion is those three sentences pointed at the island. It is close to a restatement, which is what a guaranteed completion normally looks like.
- Bferns will eventually be replaced by seed plants as the dominant vegetation.
This is real ecology, imported from outside the passage. The text is entirely about how fast reproductive units travel and who arrives first; it says nothing about dominance, competition, or what happens after arrival. A choice you can only defend with knowledge that is not on the screen is out.
- Cno plant can establish itself until moisture is present year-round.
Two inflations at once. The passage says fern spores must land somewhere moist — not that moisture must be continuous, and not that the requirement applies to every plant. "No plant" and "year-round" are both doing work the text never authorised.
- Dthe heavier a plant's reproductive structures, the less likely the plant is to survive there.
The passage links weight to how far a spore or seed travels, and therefore to when a plant arrives. Survival after arrival is a different quantity, and the text never measures it. Correct reasoning about the wrong noun.
Traps tested: Unsupported plausibility · Quantifier inflation · Wrong quantity
A first-century Roman agricultural manual advises farmers to plant vines on north-facing slopes, and historians of climate have sometimes cited the advice as evidence that first-century Italy was warmer than Italy is today: north-facing slopes receive less direct sun, which would be an advantage in a hotter climate. The same manual, however, recommends north-facing slopes for orchards in provinces whose climates it elsewhere describes as cooler than Italy's. The manual's siting advice therefore ______
Which choice most logically completes the text?
- Aconfirms that first-century Italy was warmer than Italy is today.
This is the completion the passage would have taken if it had ended before "however." The second half exists to remove exactly this reading: the manual gives the same advice for climates it calls cooler, so the advice cannot be tracking local heat. Reading to the blank without reading the connective in front of it is what leaves this choice standing.
- cannot on its own indicate how warm the climate of any particular region was.
Correct, and appropriately small. If identical advice appears for a warm region and for a region the manual itself calls cooler, then the advice does not vary with temperature, and evidence that does not vary with a quantity cannot measure it. Note what the choice declines to say: not that Italy was cool, not that the historians were wrong about the climate — only that this particular evidence cannot settle it.
- Cshows that the manual's author misunderstood the relationship between sunlight and plant growth.
The passage never evaluates whether the advice was good. There are ordinary reasons to prefer north-facing slopes that have nothing to do with a mistake — soil retention, wind, frost timing — and the text rules none of them out. A verdict on the author's competence is a claim the passage did not make.
- Dapplied only to crops that tolerate limited sunlight.
"Only" is unsupported, and so is the premise underneath it: the passage names vines and orchards without saying anything about which crops tolerate shade or whether the advice excluded others. Two facts invented in a seven-word choice.
Traps tested: Half passage completion · Unsupported plausibility · Quantifier inflation
In a recent survey of 1,200 households in a metropolitan area, 68 percent reported owning at least one motorcycle, up from 41 percent in a comparable survey conducted a decade earlier, which sampled 3,000 households. The researchers note that the earlier survey drew its sample from the entire metropolitan area, whereas the recent survey was restricted to households within five kilometres of a major arterial road. The comparison therefore ______
Which choice most logically completes the text?
- Ashows that motorcycle ownership in the metropolitan area rose by 27 percentage points.
The subtraction is right — 68 minus 41 is 27 — and that is the trap. The two percentages describe different populations: one the whole metropolitan area, the other a strip along an arterial road. Differencing them measures the change in the population sampled at least as much as any change over time. Correct arithmetic on quantities that were never comparable.
- does not establish that motorcycle ownership rose across the metropolitan area as a whole.
Correct. Because the recent sample covers only households near a major road, its 68 percent is not an estimate for the metropolitan area, and no valid comparison with the earlier area-wide figure is available. Note the restraint: the choice does not claim ownership failed to rise — it claims only that this comparison cannot show that it did, which is precisely what the sampling difference guarantees.
- Cis invalid because the recent survey included fewer households than the earlier one.
This objects to the sample size when the problem is the sampling frame. Twelve hundred households is a serviceable sample; a sample of any size drawn only from roadside households still fails to represent the metropolitan area. Naming the wrong defect leaves the real one untouched.
- Dproves that households near arterial roads own motorcycles at higher rates than other households do.
There is no contemporaneous figure for households away from arterial roads — the only such figure is ten years old and area-wide, which includes roadside households in it. With no comparison group from the same period, the difference has nothing to be a difference from, and "proves" is unavailable regardless.
Traps tested: Incomparable samples · Sample size fixation · Missing comparison group
Scholars of the Nahuatl-language annals of colonial Mexico have generally assumed that the annalists wrote for readers in their own communities, since the texts take detailed local knowledge for granted and rarely explain it. The historian Ana Delgado has proposed a further consequence: because the annalists expected their audience to know the events already, the details they chose to record indicate what those communities considered worth preserving rather than what a reader would need in order to follow the story. If Delgado's reasoning is sound, a passage in an annal that dwells at unusual length on a minor local official ______
Which choice most logically completes the text?
- Aindicates that the official's role would otherwise have been unclear to readers.
This is the reading Delgado's premise removes. Her argument starts from an audience that already knows the events, so length cannot be explained as filling a gap in the reader's knowledge — that is the one function she has ruled out. The choice contradicts the assumption the item told you to grant.
- reflects what the community regarded as worth preserving rather than what the narrative required.
Correct. Two hops, both supplied: the audience already knew the events, so selection cannot be driven by narrative necessity; and Delgado's stated consequence is that what remains, selection reflects communal valuation. Unusual length is a case of unusual selection, so the general consequence applies to it directly.
- Cshows that the annalist was personally acquainted with the official.
Plausible and unsupported. Delgado's argument is about what a community valued, which is a claim about the audience, not about the writer's social circle. A different noun, reached by a reasonable guess the passage does not make.
- Dproves that Delgado's account of the annals' audience is correct.
Circular. The prompt says "if Delgado's reasoning is sound" — you are asked to work out a consequence of an assumption, and a consequence of an assumption cannot prove the assumption. This distractor appears reliably on items that frame a hypothesis and then ask what follows from it.
Traps tested: Contradicts stated premise · Unsupported plausibility · Circular confirmation
Between 1998 and 2018 the number of peer-reviewed papers retracted each year rose roughly tenfold, a trend often cited as evidence that research misconduct is worsening. Over the same two decades, however, the number of papers published annually also rose substantially, and the share of journals with a formal retraction policy went from under 10 percent to more than 60 percent. A researcher who wanted to argue from these figures that misconduct itself has become more common would therefore need to show that ______
Which choice most logically completes the text?
- Aretractions rose faster than publication volume did, since a larger body of published work would otherwise account for the trend.
Half an answer, and the most tempting choice here. It correctly neutralises the first of the two alternatives the passage raises — more papers published — and simply omits the second. Even with the rate rising, an increase in journals capable of issuing retractions would still explain it. On items with two hedged survivors, the decision is scope, not caution.
- retractions rose faster than publication volume and that improved detection does not account for the remaining increase.
Correct. The passage supplies exactly two competing explanations for the tenfold rise — more papers, and more journals equipped to retract them — so an argument that misconduct itself increased is only available once both are ruled out. The completion names both and nothing else, which is why it is guaranteed rather than merely reasonable.
- Cmisconduct is more common in fields whose journals adopted retraction policies earliest.
A comparison the passage never sets up. Nothing in the text distinguishes fields, and a between-field difference is not among the alternatives that need eliminating — ruling it in or out would leave both stated confounds exactly where they were.
- Dthe tenfold rise in retractions exceeds any comparable increase recorded before 1998.
A historical benchmark that does nothing for the argument. Even a rise unprecedented in magnitude is still explained by more publishing and better detection if those are not addressed; the size of the number was never the objection.
Traps tested: Partial confound control · Invented comparison · Off scope benchmark
Up next
Information & Ideas — module quiz
Interleaved across all four skills — the evidence types are the confusable pair.
1 min