Reading and Writing · Information and Ideas

SAT Inferences: The Complete Guide

Inference questions give you a short passage with a missing final sentence or conclusion and ask what logically follows. The right answer goes one small step beyond the text and no further. This guide teaches you to find that step, to recognize how the wrong answers overreach or drift, and to practice the habit with examples.

Section: Reading and Writing
Domain: Information and Ideas (about 26% of the section)
Questions per test: Roughly 3–5
Typical stem: “Which choice most logically completes the text?”

What Inferences tests

An inference is a conclusion that the text makes very likely but does not state outright. On the SAT, inference questions present a short passage ending in a blank, usually where a conclusion would go, and ask which choice most logically completes it. Sometimes the question asks what the text suggests, implies, or allows you to conclude.

The passages are drawn from science, social science, humanities, and literature, but the logic is the same: the passage gives you premises (observations, findings, background), and you choose the conclusion that the premises most strongly support. It is a test of disciplined reasoning, not imagination.

Inferences differ from central idea questions in one key way. Central idea questions ask what the text says; inference questions ask what the text lets you conclude. The answer is new information, but only information the premises already force.

The method at a glance: premises, prediction, match

Read the passage and list the premises: the facts, findings, or observations. Then ask what must be true, or is very likely true, given all of them, and especially given the last sentence before the blank. Write the conclusion in your own words before looking at the choices. Choose the answer closest to your prediction, and reject any that add information, upgrade a hedge, or answer a different question.

In nearly every inference passage, the final sentence before the blank contains the key evidence. Earlier sentences usually set up context. If you only have time to use one sentence, use the one right before the blank.

Core concepts: what a good inference looks like

Valid versus invalid inferences
Passage gives	Valid inference	Invalid inference
Plants without wax lost water faster	Wax likely helps plants conserve water	Wax is the only thing that helps plants conserve water
Pottery clay matches a distant deposit	The community likely obtained clay or pottery through some contact	The community traveled to the deposit by ship
Sales fell more where competitors were close	Nearby alternatives may affect how customers respond to prices	Competition always reduces sales

Hedging. Valid inferences are usually hedged: “likely,” “may,” “suggests.” This is because short passages rarely prove anything. When a choice states something absolutely and a hedged choice says almost the same thing, the hedged choice is usually the better one.

Scope. The inference must stay within the group, place, or time that the passage discusses. A study of one lake supports statements about that lake, not all lakes. A finding about adults does not apply to children.

Causal language. A correlation in the passage supports “may be related to,” not “causes.” If the passage explicitly rules out other causes or runs a controlled comparison, causal language becomes safer.

Step-by-step method

Read the whole passage and list the premises. What did the researchers observe, find, or claim?
Find the final clue. The sentence right before the blank, and the logical connector, tell you what the blank must do.
Predict the conclusion in your own words. Keep it short and hedged.
Match choices to your prediction. Prefer the choice that stays closest to the premises.
Test your top choice. Ask: could I defend this to a skeptic using only the passage? Does it avoid words like “always,” “only,” and “proves”?
Eliminate by naming the flaw. Overreach, scope shift, reversal, or irrelevance.

When two choices both seem valid, one of them nearly always relies on an additional assumption. The one that needs fewer assumptions is the answer.

Worked examples

Worked example

Example 1 (easy): a conclusion from an experiment

Many desert plants have a waxy coating on their leaves. A botanist carefully removed the coating from the leaves of several cactus seedlings and left the leaves of an identical set of seedlings untouched. Over the next week, the seedlings without the coating lost water much faster than the untouched seedlings. This result suggests that ______

Which choice most logically completes the text?

Athe waxy coating helps desert plants conserve water.
Bcactus seedlings grow faster without a waxy coating.
Cwaxy coatings are found only on desert plants.
Dthe botanist’s experiment worked on every plant species.

Show the solution

Premises: two identical groups, one without the coating. The coating-free group lost water faster.
Prediction: the coating probably slows water loss.
Choice B introduces growth, which the passage never measured.
Choice C overreaches. The passage says many desert plants have wax, not that only desert plants do.
Choice D makes a claim about every plant species. The experiment used cactus seedlings only, a scope shift.
Choice A matches the prediction and is hedged by “helps.”

Answer: A

Worked example

Example 2 (medium): a hedged inference from evidence

Archaeologists excavating a coastal village found pottery fragments made from a clay whose chemical signature matches deposits located about 300 miles inland. No clay of that type has been found anywhere near the village. The researchers note that the pottery shows no signs of being a later addition and appears to date from the earliest layers of the settlement. This evidence suggests that ______

Which choice most logically completes the text?

Athe villagers had some form of contact with communities near the inland deposits from an early stage.
Bthe villagers had invented pottery before any inland communities did.
Cthe villagers regularly traveled by sea to collect clay from the inland deposits.
Dthe inland communities were more technologically advanced than the villagers.

Show the solution

Premises: coastal pottery uses clay matching distant inland deposits; no such clay nearby; the pottery is from the earliest layers.
Prediction: the villagers got the clay or the pots from somewhere inland, probably through contact or trade, from early on.
Choice B is about who invented pottery first. Nothing in the passage compares invention dates.
Choice C is too specific. The deposits are inland, so traveling “by sea” to collect clay is unsupported, and “regularly” is not established either.
Choice D makes a claim about technological advancement with no evidence.
Choice A is hedged and stays within the premises: the clay’s source implies contact, and the early dating supports “from an early stage.”

Answer: A

Worked example

Example 3 (hard): a conditional inference

Economists studied how two retail chains’ sales responded to the same 5 percent price increase. At Chain A, sales did not change. At Chain B, sales fell by 8 percent. The researchers observed that Chain A’s stores are mostly located in areas with no nearby competitors, whereas Chain B’s stores typically have several competing stores within a short drive. These results suggest that ______

Which choice most logically completes the text?

Araising prices always reduces sales at stores that have nearby competitors.
Bhow much a price increase affects sales may depend on whether customers have alternatives.
CChain A sells products of higher quality than the products Chain B sells.
Dcustomers who live far from competing stores do not notice price changes.

Show the solution

Premises: same price increase; no sales change where competitors are absent; a sales drop where competitors are near.
Prediction: the presence of alternatives may shape the sales response to a price rise.
Choice A is overreach. “Always” cannot come from one comparison of two chains.
Choice C introduces quality, which the passage never mentions. It is an alternative explanation, not a conclusion supported by the data.
Choice D makes a claim about what customers notice. The passage reports sales, not awareness, so this is a plausible but unsupported reading.
Choice B states a hedged, conditional relationship (“may depend on”) that fits both chains’ results and the researchers’ observation.

Answer: B

Common traps and how to beat them

What separates a 700 from a 600

Hard inference questions use dense or abstract premises, so you must do more work to see what follows. The structure of the wrong answers is also more subtle.

Multiple premises must be combined. The right conclusion uses two or more sentences together. A choice built on only one sentence is often a distractor.
Conditional answers. The best choice may be “depends on” or “only under certain conditions,” which feels less satisfying than a firm claim but is better supported.
Concession and contrast. If the passage says “although” or “despite,” the conclusion must respect both sides. Answers that ignore the concession are wrong.
Almost-valid logic. Some wrong choices are logical for a slightly different premise. Compare each choice with the exact words of the passage, not the gist.
Humanities and literature. Inferences about a character’s state of mind or an author’s attitude must be anchored in specific words or actions, not general impressions.

Timing and a practice plan

Inference questions often take 70 to 90 seconds because predicting the conclusion takes thought. Do not skip the prediction to save time; it is what makes elimination fast.

Days 1–2: Do 8 inference questions untimed. For each, write your predicted conclusion before reading choices, then compare it to the correct answer.
Days 3–4: Do 8 questions at about 80 seconds. For every wrong choice, label the flaw: overreach, scope shift, new information, correlation as cause, or reversal.
Days 5–6: Do a mixed set that includes hard questions. Practice ranking choices by how many assumptions each needs.
Day 7: Review your error log and rewrite three of your misses as stronger, hedged conclusions.

Keep reading — it’s free

Create a free Solar account to unlock the rest of this inferences guide, plus adaptive practice on exactly this skill.

Core concepts: what a good inference looks like
Step-by-step method
Worked examples
Common traps and how to beat them
What separates a 700 from a 600

Unlock the full guide Already have an account? Log in

Inferences: FAQ

How many inference questions are on the digital SAT?

Expect roughly 3 to 5 per test across the Reading and Writing section. The exact number varies from test to test.

What is an inference on the SAT?

An inference is a conclusion the passage strongly supports but does not state outright. It stays one small logical step from what the text says.

Why are hedged answers often correct on SAT inference questions?

Short passages rarely prove anything, so the best-supported conclusion is usually hedged with words like “may,” “suggests,” or “likely.” Absolute words tend to overreach.

How do I handle inference questions with unfamiliar science topics?

Ignore the jargon you do not know and focus on the structure: what was compared, what changed, and what result followed. The logic does not depend on the vocabulary.

How can I avoid choosing answers that sound right but are not supported?

Predict the conclusion before reading the choices and test each choice by asking whether you could defend it to a skeptic using only the passage.