Reading from First Principles

From-zero teaching for the reading skills — Inferences & Command of Evidence first

49 min read

Built for a strong math brain that wants Reading to feel like a proof, not a vibe. Everything below is original teaching material with original examples. When you want real, official practice, use the College Board Question Bank (Bluebook) — that is the only source that is guaranteed to match the live test.

#How this guide is organized

Part 0 — How to actually read a digital-SAT passage (the mindset that fixes "inconsistent").
Part 1 — Inferences (the hardest, most-missed type). Deep dive.
Part 2 — Command of Evidence, Textual. Deep dive.
Part 3 — Command of Evidence, Quantitative (tables/graphs). Deep dive.
Part 4 — The remaining Reading types, fast.
Part 5 — A universal checklist and the exact order to learn and drill.

A note on the format you're facing: the digital SAT gives you one short passage (about 25–150 words) with exactly one question. There is no long article with 10 questions anymore. Each item is self-contained. This is great news for a logical thinker — every question is a small, closed puzzle where the answer is inside the box.

#PART 0 · HOW TO READ AN SAT PASSAGE (THE MINDSET)

#The single idea that fixes inconsistency

The correct answer is PROVABLE using only the words on the screen. Every wrong answer requires you to ASSUME something, ADD something, or IMPORT outside knowledge.

Treat this like math. In a proof you may only use the given axioms and what follows from them. On SAT Reading, the passage is your set of axioms. If a choice needs a fact that isn't given — even a "true" fact you happen to know — it is out of bounds. Reading gets inconsistent for smart students precisely because they know too much: they fill gaps with real-world knowledge the passage never granted them. Stop doing that and your accuracy stabilizes.

#Why "strong at math, weak at reading" is usually a process problem, not a talent problem

Math rewards a fixed algorithm. Most students have no algorithm for Reading — they read, "feel" an answer, and pick. That's why the same student gets 5 in a row then misses 3 easy ones. The cure is to run the same five steps every single time, even on questions that feel obvious. Consistency comes from process, not from reading faster or "getting" English better.

#The 5-step read (run this on EVERY question)

Step 1 — Read the question stem FIRST. Before the passage, glance at what's being asked. "Which choice best states the main idea?" reads differently from "Which finding, if true, would weaken the researchers' claim?" Knowing the job changes what you hunt for.

Step 2 — Read the whole passage, slowly, once. It's short. Do not skim. Find the point: what is the author actually claiming or doing? In a 60-word passage there is usually one main claim and one or two supporting sentences. Say the point to yourself in plain words.

Step 3 — Mark the logic words. These are the joints of the argument. They tell you how ideas connect:

Logic word	What it signals
however, but, yet, although, whereas, despite, on the other hand	a turn — the idea is about to reverse or contrast
therefore, thus, so, as a result, consequently	a conclusion — what follows is the point/payoff
because, since, given that, due to	a reason/cause
for example, for instance, such as	an illustration of a bigger claim
moreover, furthermore, in addition	more of the same direction

The word right after "however" or "therefore" is very often where the answer lives. Circle these mentally.

Step 4 — PREDICT before you read the choices. In your own words, answer the question before looking at A, B, C, D. This is the most important habit on the entire section. If you predict "the author thinks the new method is promising but unproven," you now have a template. Choices that don't match your prediction get eliminated fast, and you're far less likely to be seduced by a slick wrong answer. Prediction is your defense against traps.

Step 5 — Eliminate with text evidence. For each choice ask: "Can I point to the specific words that prove this?" If you can't put your finger on the proof, it's not the answer. Do not argue for a choice with your imagination; argue against three choices with the text. The last one standing wins.

#The anatomy of a wrong answer (memorize these five)

Almost every trap answer is one of these. Learning to name the trap is how you kill it under time pressure.

Overreach / too extreme. Adds "always," "never," "only," "proves," "impossible." The passage said "suggests"; the choice says "proves." Out.
Outside knowledge / true-but-unsupported. A real fact about the world that the passage never stated. True ≠ supported. Out.
Half-right (the wrecker). The first half matches; the second half sneaks in a distortion. Read every word of every choice — SAT hides the error at the end.
Opposite / reversed. Correct topic, wrong direction. Says "increased" when the text says "declined," or blames the author's view on the people the author is criticizing.
Off-topic / plausible but absent. Sounds reasonable and on-theme but answers a question that wasn't asked, or discusses something the passage didn't.

Whenever you're stuck between two choices, ask which one commits one of these five sins. Usually one does, cleanly.

#PART 1 · INFERENCES (THE HARDEST TYPE)

#What an inference question is

The stem looks like:

"Which choice most logically completes the text?" "Based on the text, what can most reasonably be concluded / inferred?" "Which statement about ___ is best supported by the text?"

Your job: state what must be true — or what the passage is clearly steering toward — given only the sentences provided. Many of these end with a blank ("_______") that you complete, or a "logically completes" phrasing. Same skill either way.

#The mental model: this is deduction, not brainstorming

Think of the passage as premises in a logic problem. The right inference is the smallest, safest step that the premises force. It is not the most interesting idea, not the most sophisticated, not the one that shows off what you know. It's the conclusion you'd be embarrassed to deny after reading the text.

The test to apply to every choice: "Given ONLY these sentences, does this HAVE to be true (or is it the only thing that makes the passage cohere)?" If you can imagine the passage being true while the choice is false, the choice is not a valid inference.

#The four ways students overreach on inferences (name them, avoid them)

Overreach. The text supports a mild claim; you pick a strong one. Text: "attendance rose in two of the five years studied." Bad inference: "attendance is rising." (Two of five isn't a trend.)
Prediction of the future. The text describes a present state; you infer what "will" happen. Unless the text gives a trend or a stated cause-and-effect, the future is off-limits.
Correlation treated as causation. Two things occur together; you infer one caused the other. The SAT loves this trap. "Sales rose the same month the ad ran" does not license "the ad raised sales."
Outside knowledge. You know something true about the topic and quietly use it. Forbidden. Only the given words count.

The correct inference dodges all four: it stays mild, stays in the present (unless a trend is stated), claims no causation beyond what's stated, and uses only given facts.

#Worked Example 1 (warm-up)

A city installed motion-sensing streetlights that dim to 40% brightness when no one is nearby and return to full brightness when they detect a pedestrian or vehicle. In the first year, the district using these lights reported the same number of nighttime traffic incidents as it had the previous year, while its street-lighting electricity use fell by nearly a third.

Which choice most logically completes the text? The results suggest that, in this district, the new lights ______

A) eliminated nighttime traffic incidents entirely. B) reduced electricity use without a measurable increase in nighttime traffic incidents. C) were more popular with residents than the old lights. D) will soon be adopted by every district in the city.

Predict first. The text gives two facts: incidents stayed the same, electricity dropped ~1/3. A safe combined statement: "saved energy without making traffic incidents worse." Now match.

A — Overreach/Opposite. "Eliminated entirely" contradicts "the same number of incidents" (which wasn't zero unless stated, and "same as previous year" implies incidents still occurred). Out.
B — Correct. "Reduced electricity use" = the one-third drop. "Without a measurable increase in incidents" = "same number." Every piece is provable. This is the mild, forced conclusion.
C — Outside knowledge. Popularity with residents is never mentioned. Might be true in real life; unsupported here. Out.
D — Prediction of the future. Citywide adoption is a forecast the text gives zero basis for. Out.

Trap-level thinking vs. correct thinking: A hurried reader sees "streetlights good" and grabs A ("entirely!") because it sounds like a strong win. The disciplined reader notices entirely is stronger than the text ("same number," not "zero") and rejects it. The answer is deliberately unexciting. B.

#Worked Example 2

Marine biologist Dr. Okafor studied a reef where a certain cleaner shrimp removes parasites from fish that visit it. She noticed that fish lined up and waited their turn at stations where these shrimp lived, sometimes passing up feeding opportunities to do so. Fish did not wait at otherwise identical rock formations that lacked the shrimp.

Based on the text, what can most reasonably be inferred about the fish's waiting behavior?

A) The fish are incapable of removing parasites without assistance. B) The fish's willingness to wait is connected to the presence of the cleaner shrimp rather than to the rock formations themselves. C) The cleaner shrimp prefer certain species of fish over others. D) Reefs without cleaner shrimp have unhealthier fish populations.

Predict. The controlled contrast is doing the work: fish wait where shrimp are, don't wait at identical rocks without shrimp. So the shrimp, not the rocks, drive the waiting. That's the safe inference.

A — Overreach + outside knowledge. "Incapable" is far too strong; the text says fish choose to wait, not that they can't self-clean. Out.
B — Correct. The "otherwise identical" rocks without shrimp are the key: they isolate the shrimp as the variable. This is textbook controlled-comparison reasoning, and every word is anchored. In.
C — Off-topic. Shrimp preferences among fish species are never discussed. Out.
D — Overreach/future/outside. "Unhealthier populations" is a big causal claim about reef health the passage never measures. Out.

Trap contrast: The tempting wrong answer is D — it feels like the "deeper meaning" (shrimp help fish, so no shrimp = sick fish). But that's you writing the study's conclusion for it. The passage only measured waiting behavior, so your inference must stay about waiting behavior. Notice how the "otherwise identical" phrasing is a gift: the SAT plants controlled-comparison language exactly so that the disciplined answer is the one that points to the isolated variable. B.

#Worked Example 3 (mid-difficulty, a causation trap)

An economist reviewed data from a coastal town after it opened a large public aquarium downtown. In the three years following the opening, downtown restaurant revenue grew faster than it had in the three years before. The economist cautioned, however, that the region's overall tourism also rose sharply during the same period for unrelated reasons.

Which choice most logically completes the text? The economist's caution implies that ______

A) the aquarium had no effect on restaurant revenue. B) the rise in restaurant revenue cannot be attributed to the aquarium with confidence, because another factor changed at the same time. C) downtown restaurants would have failed without the aquarium. D) rising regional tourism was caused by the new aquarium.

Predict. Watch the logic word however — it introduces a caution. Two things changed at once (aquarium opened and tourism rose). So you can't credit the aquarium alone. The safe inference: "we can't confidently say the aquarium caused the revenue growth."

A — Overreach/Opposite. "No effect" is as unsupported as "definitely caused it." The economist says we can't be sure, not that there was zero effect. Out. (Classic: uncertainty ≠ proof of no effect.)
B — Correct. This is precisely what a confound does — a second simultaneous variable prevents a confident causal claim. Mild, exact, provable. In.
C — Overreach. "Would have failed" is a dramatic counterfactual with no support. Out.
D — Reversed causation + unsupported. The text says tourism rose "for unrelated reasons," explicitly denying the aquarium caused it. Out.

Trap contrast: A math-strong student who's sloppy with English might over-correct into A: "if we can't prove it helped, then it didn't help." That's a logic error — absence of evidence isn't evidence of absence. B captures the real epistemics: we simply can't isolate the cause. This is the single most important inference pattern on the SAT — two variables change together, so no causal claim is safe. Learn it cold. B.

#Worked Example 4 (harder — subtle scope)

In a study of memory, participants who took a short walk outdoors between studying a word list and being tested on it recalled more words than participants who spent the same interval sitting indoors. A second group who walked indoors on a treadmill for the same interval recalled roughly as many words as the outdoor walkers, and more than those who sat.

Based on the text, which conclusion is best supported?

A) Being outdoors, rather than physical movement, is what improved recall. B) Light physical activity during the interval, whether indoors or outdoors, was associated with better recall than sitting. C) Walking improves memory permanently. D) Sitting indoors harms the ability to form new memories.

Predict. Compare the three groups: outdoor walkers ≈ indoor treadmill walkers > sitters. What's common to both high-recall groups? Walking (movement), not the outdoors — because indoor walkers did just as well. So movement, not setting, tracks with recall.

A — Reversed / contradicted. The indoor treadmill group disproves "outdoors is the key," since they matched the outdoor group without being outdoors. This choice picks the wrong variable. Out.
B — Correct. The two walking groups (indoor and outdoor) both beat sitting, and matched each other — so the shared factor is activity, and the honest verb is "was associated with" (not "caused"). Perfectly scoped. In.
C — Overreach + future. "Permanently" is enormous; the study measured one test session. Out.
D — Overreach/Opposite framing. "Harms the ability to form memories" is a strong causal-damage claim; the data only show sitters recalled fewer, not that sitting harms memory formation. Out.

Trap contrast: A reader who fixates on the vivid detail ("outdoors!") grabs A. But the second group is placed there specifically to rule out the outdoors. The skill: when a passage adds a second comparison group, ask "what does this new group let me rule out?" Here it rules out "outdoors" and points at "movement." Also note B's careful "was associated with" — the SAT rewards the choice that refuses to overclaim causation. B.

#Worked Example 5 (hardest — "logically completes" with a turn)

Historians once assumed that a medieval town's wealth could be estimated from the size of its central church, reasoning that only prosperous towns could fund grand construction. But recent analysis of tax records shows that several towns with modest churches were among the wealthiest in their regions, while some towns with soaring cathedrals carried heavy debts from the building projects themselves. This evidence indicates that ______

A) church size is a reliable measure of a medieval town's wealth. B) a town's church size is not a dependable indicator of its actual wealth. C) medieval towns should not have built large cathedrals. D) tax records are the only trustworthy source for medieval history.

Predict. The logic word But flips the old assumption. New evidence: rich towns had small churches; some cathedral-towns were in debt. So church size and wealth don't track reliably. The blank should say: "church size doesn't reliably indicate wealth."

A — Opposite. This is the old assumption the "But" overturns. Out.
B — Correct. Both pieces of evidence break the church-size↔wealth link, so the size is "not a dependable indicator." Directly forced by the text. In.
C — Outside/value judgment. "Should not have built" is an opinion the text doesn't make; the passage is about measuring wealth, not about whether building was wise. Out.
D — Overreach. "The only trustworthy source" is a sweeping claim; tax records are used here, not crowned the sole reliable source. Out.

Trap contrast: After a contrast structure, the sloppy move is to grab the sentence that sounds smart (C's editorializing, or D's grand claim about sources). But "logically completes" wants the conclusion the evidence forces — and the evidence is narrowly about whether church size predicts wealth. Stay on the exact variable the evidence addresses. When you see But / However flip an old belief, the completion is almost always "the old belief is unreliable/wrong," stated mildly. B.

#Inference — Mini-Practice (8 items)

Predict before looking at choices. Name the trap in each wrong answer as you eliminate.

P1. A bakery switched from plastic to compostable packaging. In the following quarter, its packaging costs rose, but it also gained several new customers who said in a survey that they chose the bakery specifically because of the eco-friendly packaging. Which most logically completes the text? The switch ______

A) lowered the bakery's total expenses.
B) increased packaging costs while attracting at least some customers who valued the change.
C) was opposed by most existing customers.
D) guaranteed the bakery's long-term success.

P2. Researchers found that a species of frog calls more loudly near noisy streams than near quiet ponds. Frogs relocated from quiet ponds to noisy streams increased their call volume within days. Best supported inference?

A) The frogs' call volume can adjust to the noise level of their surroundings.
B) The frogs prefer streams to ponds.
C) Loud calling damages the frogs' hearing.
D) All frog species adjust their calls to noise.

P3. A company let employees choose whether to work from home. Productivity, measured by projects completed, stayed about the same as the year before, but reported employee satisfaction rose. Which most logically completes the text? The policy ______

A) reduced the number of projects completed.
B) was associated with higher satisfaction without a drop in projects completed.
C) proved that home work is better for every company.
D) made employees less committed to their jobs.

P4. An archaeologist noted that a certain style of pottery appears at coastal sites but not at inland sites from the same period, even though both regions traded in other goods. Best supported inference?

A) The pottery style was not distributed to inland sites during that period.
B) Inland people disliked the pottery.
C) The coastal potters were more skilled.
D) The inland sites are younger than the coastal sites.

P5. A tutoring program reported that students who attended at least ten sessions improved their grades more than students who attended fewer. However, students chose for themselves how many sessions to attend. Which most logically completes the text? The results ______

A) prove that attending more sessions causes higher grades.
B) cannot by themselves establish that the extra sessions caused the greater improvement, since more-motivated students may have chosen to attend more.
C) show the program had no effect.
D) show that grades determine attendance.

P6. A newly described deep-sea fish has eyes far larger, relative to its body, than those of related fish living in shallower water. Best supported inference?

A) The fish's large eyes may be related to living where little light reaches.
B) The fish is blind.
C) The fish is the largest in its family.
D) Shallow-water fish have poor vision.

P7. A town replaced its bus fleet with electric buses. Air-quality sensors downtown recorded lower levels of certain pollutants the following year. City officials noted that a nearby factory, a major polluter, had also closed that year. Which most logically completes the text? The improvement in air quality ______

A) was caused entirely by the electric buses.
B) cannot be confidently credited to the buses alone, because another major source of pollution also changed.
C) shows the factory was harmless.
D) will continue indefinitely.

P8. In a taste test, participants rated two identical soups differently when told one was "homemade" and the other "canned," giving higher ratings to the one labeled "homemade." Best supported inference?

A) The label given to a food can influence how people rate its taste.
B) Homemade soup tastes better than canned soup.
C) The participants had unusually sensitive palates.
D) Canned soup is unhealthy.

#Inference — Answer Key

P1 — B. A = opposite (costs rose). C = unsupported (no data on existing customers' opposition). D = overreach ("guaranteed").
P2 — A. Relocated frogs changed volume → volume tracks environment noise. B = off-topic (preference). C = outside/unsupported. D = overreach ("all species").
P3 — B. "Same" projects + higher satisfaction. A = opposite. C = overreach ("every company"). D = unsupported/opposite.
P4 — A. Present only at coastal sites = the safe, literal statement. B = mind-reading motive. C = outside (skill). D = unsupported dating claim.
P5 — B. Self-selection = confound → no causal claim. A = overreach (causation). C = opposite. D = reversed and unsupported.
P6 — A. Mild "may be related to little light" — safe. B = overreach ("blind"). C = off-topic (size). D = overreach about shallow fish.
P7 — B. Two pollution sources changed together → confound. A = overclaim. C = unsupported reversal. D = future overreach.
P8 — A. The only variable was the label → label influences ratings. B = opposite (soups were identical!). C = unsupported. D = outside knowledge.

#PART 2 · COMMAND OF EVIDENCE — TEXTUAL

#What this type asks

You get a claim, a hypothesis, an argument, or a student's conclusion — then:

"Which finding, if true, would most directly support the claim?" "Which quotation most effectively illustrates the argument?" "Which finding would most weaken / undermine / call into question the hypothesis?" "Which additional detail, if true, would most strongly support the researcher's conclusion?"

You are handed a target and asked which piece of evidence hits it. This is inductive reasoning: evidence makes a conclusion more or less likely (it doesn't prove it with certainty the way an inference is forced). Your job is to judge relevance and direction.

#The method (four moves)

Move 1 — Pin the EXACT claim. Write it in your head as a crisp sentence with a subject, a variable, and a direction. Vague claim-holding is where students lose these. Example claim: "This fungus helps its host tree by delivering nitrogen to the tree's roots." The variable is nitrogen delivered to the tree; the direction is helps the tree.

Move 2 — Read the TASK VERB and obey it. Support, weaken, and illustrate are different jobs:

Support → find evidence that, if true, makes the claim more likely.
Weaken/undermine → find evidence that, if true, makes the claim less likely (often by breaking its logic or providing a rival explanation).
Illustrate/exemplify → find the concrete example that fits the general claim (no proving required — just a clean instance).

Students who "support" when the verb says "weaken" auto-fail. Underline the verb.

Move 3 — Ask "what would this claim PREDICT?" If the claim were true, what should we observe in the world? For the fungus claim: trees with the fungus should have more nitrogen / grow better than trees without it. Now you know what supporting evidence looks like before you read the options.

Move 4 — Test each choice against the exact variable AND direction. A choice must (a) be about the same variable and (b) push the right way. Right topic but wrong direction = trap. Right direction but different variable = trap.

#The traps specific to textual evidence

Right topic, wrong direction. Evidence about the same subject that actually cuts the other way. (For a "support" question, a choice that would weaken the claim — and vice versa.)
Irrelevant / different variable. True and on-theme, but about a different quantity than the claim. (Claim is about nitrogen; choice is about the fungus's color.)
Too weak / restates the claim. Doesn't add evidence, just repeats the assertion or gives an unrelated fact.
Over-broad. Evidence about a different population or a bigger group than the claim covers.
Real-world plausible but off-target. Sounds like something that "should" support it but doesn't address the specific logic.

Golden test for textual evidence: "If this choice were TRUE, would it push the claim in the direction the task verb asks — using the claim's exact variable?" Only a yes on both counts survives.

#Worked Example 1 (support)

A botanist hypothesizes that a particular desert shrub survives droughts by growing an unusually deep root system that reaches moisture far below the surface, rather than by storing water in its stems as some desert plants do.

Which finding, if true, would most directly support the botanist's hypothesis?

A) The shrub's stems contain less stored water than those of most other desert plants of similar size. B) During a severe drought, the shrub's roots were found to extend several meters deeper than those of neighboring plants that died. C) The shrub produces bright flowers after seasonal rains. D) The shrub grows more slowly than many other desert plants.

Pin the claim + predict. Claim: the shrub survives droughts via deep roots reaching deep moisture, not via stem storage. Task verb: support. Prediction: we'd want evidence that its roots go unusually deep and that this tracks with surviving drought.

A — Right topic, but supports only half — and weakly. Low stem water rules out stem-storage, which is consistent, but it doesn't show the deep-root mechanism that is the heart of the claim. It's the runner-up, not the answer: it addresses the "not by storing water" clause but not the load-bearing "deep roots reach moisture and enable survival."
B — Correct. Roots "several meters deeper" (deep root system ✓) plus the contrast with "neighboring plants that died" during drought (the roots track with surviving the drought ✓). Hits the exact variable and direction. In.
C — Irrelevant. Flowering after rain says nothing about drought-survival mechanism. Out.
D — Irrelevant. Growth rate is a different variable. Out.

Trap contrast: A is engineered to catch students who grab the first on-topic choice. It's true and relevant but only knocks down the alternative theory; it never demonstrates the deep-root mechanism doing the drought-surviving work. B supplies both the mechanism (depth) and the payoff (survival vs. death). When two choices both look relevant, pick the one that supports the main mechanism, not just a side clause. B.

#Worked Example 2 (weaken)

A researcher claims that a new phone app improves users' sleep, citing that people who downloaded the app reported sleeping longer after two weeks of use.

Which finding, if true, would most weaken the researcher's claim?

A) The app is inexpensive and easy to use. B) People who chose to download a sleep app were already planning lifestyle changes to sleep more, which they carried out regardless of the app. C) The app has been downloaded by millions of users. D) Some users found the app's reminders annoying.

Pin the claim + predict. Claim: the app causes longer sleep. Task verb: weaken. The claim's weak point is causation (self-reported, self-selected users). To weaken, provide a rival explanation for the longer sleep that isn't the app.

A — Irrelevant. Price/ease doesn't bear on whether it works. Out.
B — Correct. If downloaders were already changing habits and would have slept more anyway, then the app isn't the cause — the improvement is explained by something else. This directly undercuts the causal claim. In.
C — Irrelevant (and arguably wrong direction). Popularity isn't effectiveness; millions of downloads doesn't tell us the app improves sleep. Out.
D — Too weak / different variable. Annoying reminders is about satisfaction, not about whether sleep improved. Out.

Trap contrast: Students hunting to "weaken" often grab any negative-sounding detail (D — "annoying!"). But weakening a causal claim means offering a confound or alternative cause, not just saying something bad about the product. B is the alternative-cause bomb. The pattern to memorize: to weaken a causal claim, find the rival explanation. B.

#Worked Example 3 (illustrate)

In an essay on economical writing, a critic argues that skilled authors often convey a character's entire personality through a single, well-chosen action rather than through long descriptions.

Which quotation from a novel would most effectively illustrate the critic's claim?

A) "Mara was, by every account, a generous, impatient, and endlessly curious woman who had traveled to more countries than she could name." B) "Without breaking stride, Mara pressed her last coin into the street musician's palm and was gone before he could thank her." C) "The city stretched out before her, gray and immense, humming with the noise of a thousand lives." D) "Mara had always believed that kindness was the highest virtue, a principle she discussed often with her friends."

Pin the claim + predict. Claim: personality shown through a single action, not long description. Task verb: illustrate — find the clean example of that technique. Prediction: a quote where one small action reveals character, with no explaining.

A — Opposite technique. This is exactly the long description the critic contrasts against ("generous, impatient, endlessly curious…"). It tells, doesn't show. Out.
B — Correct. One action — pressing her last coin on a musician and slipping away — shows generosity and modesty without a word of description. This is the technique in action. In.
C — Different variable. Vivid, but it describes the city, not a character's personality. Out.
D — Opposite technique. Character conveyed through stated beliefs and discussion — description/telling, not a single revealing action. Out.

Trap contrast: For "illustrate," don't grade the writing on how pretty it is (C is the prettiest and it's wrong). Grade it on whether it matches the described technique. B is the only quote where a single action carries the characterization. Match the method, not the mood. B.

#Worked Example 4 (support — data-flavored but still textual)

A team argues that a wild songbird learns its song primarily by imitating adult birds it hears while young, rather than by inheriting the song genetically.

Which finding, if true, would most directly support the team's argument?

A) Young birds raised in isolation, hearing no adult songs, grew up to sing songs unlike those of their own species. B) Adult birds of this species sing more often in spring. C) The species' song is more complex than that of closely related species. D) Birds of this species have a wider vocal range than most songbirds.

Pin the claim + predict. Claim: song is learned by imitation, not inherited. Task verb: support. Prediction: if learning matters, birds deprived of adult models should fail to develop the normal song. (This is the classic deprivation experiment — remove the proposed cause, see if the effect disappears.)

A — Correct. Isolated birds (no adult song to imitate) develop abnormal song → learning by imitation is necessary, and it's not purely genetic. Exactly targets the learned-vs-inherited variable. In.
B — Irrelevant. Timing of singing says nothing about learning vs. inheritance. Out.
C — Irrelevant. Complexity relative to other species doesn't address the mechanism of acquisition. Out.
D — Irrelevant. Vocal range is a different variable. Out.

Trap contrast: B, C, and D are all true-sounding bird facts — the "on-theme but off-target" trap. Only A engages the actual claim's logic (learning requires exposure). The deep move: to support a "learned, not inherited" claim, look for the choice that removes the learning opportunity and sees the trait vanish. That's a designed experiment doing exactly what the claim predicts. A.

#Textual Evidence — Mini-Practice (6 items)

P1. Claim: A city planner argues that adding protected bike lanes on a busy street made cyclists safer. Which finding, if true, would most support the claim?

A) The number of cyclists using the street increased after the lanes were added.
B) After the lanes were added, collisions involving cyclists on that street fell substantially, even as cycling numbers stayed steady.
C) The bike lanes were painted a bright, visible color.
D) Drivers reported that the street looked more crowded.

P2. Claim: A nutritionist claims that a breakfast high in protein keeps people feeling full longer than a breakfast high in sugar. Which finding would most weaken the claim?

A) In a controlled test, people who ate the high-protein breakfast reported feeling hungry again just as soon as those who ate the high-sugar breakfast.
B) High-protein breakfasts are more expensive.
C) Some people prefer sweet breakfasts.
D) Protein takes longer to digest than sugar.

P3. Claim: A literature teacher argues that a poet uses images of winter to suggest emotional isolation. Which quotation would best illustrate the claim?

A) "The summer fair was loud with laughter and the smell of ripe fruit."
B) "I walked alone across the frozen field, my only company the wind and the bare, silent trees."
C) "She counted the coins twice before handing them over."
D) "The poem is written in careful, regular meter."

P4. Claim: Researchers hypothesize that a lizard changes color mainly to regulate its body temperature, not to communicate with other lizards. Which finding would most support the hypothesis?

A) The lizard turns darker on cold mornings and lighter in the midday heat, even when no other lizards are present.
B) The lizard is more colorful during mating season.
C) The lizard lives in a wide range of habitats.
D) Other lizard species change color to signal aggression.

P5. Claim: A historian argues that a printing innovation spread literacy by making books cheaper. Which finding would most support the claim?

A) In regions that adopted the innovation, book prices dropped and the share of people who could read rose in the following decades.
B) The innovation was invented by a single craftsman.
C) Books printed with the innovation were physically smaller.
D) The innovation was later replaced by newer technology.

P6. Claim: A scientist claims that a coral species survives warming water better than others because it hosts a heat-tolerant type of algae. Which finding would most weaken the claim?

A) When researchers removed the heat-tolerant algae, the coral survived warming just as well as before.
B) The coral is brightly colored.
C) The heat-tolerant algae are also found in other coral species.
D) The coral grows slowly.

#Textual Evidence — Answer Key

P1 — B. Fewer collisions with steady numbers = safer per cyclist. A = wrong variable (more riders ≠ safer). C, D = irrelevant.
P2 — A. Same variable (fullness/hunger), opposite direction → weakens. D actually supports. B, C = irrelevant.
P3 — B. Winter imagery (frozen field, bare trees) + isolation ("alone"). A = wrong season/mood. C, D = off-topic.
P4 — A. Color tracks temperature and happens with no other lizards present — rules out communication. B = supports the rejected communication idea. C, D = irrelevant.
P5 — A. Price drop + literacy rise = the claimed mechanism and effect. B, C, D = irrelevant to the price→literacy link.
P6 — A. Remove the proposed cause (algae), effect persists → the algae aren't why. B, D = irrelevant. C = doesn't address this coral's survival.

#PART 3 · COMMAND OF EVIDENCE — QUANTITATIVE

#What this type asks

You get a table, bar graph, or line graph, plus a short passage describing a study, and a stem like:

"Which choice most effectively uses data from the graph to support the claim?" "Which choice best describes data from the table that support the student's conclusion?" "Which choice, if true, would complete the text with data consistent with the figure?"

You don't need math skill here. You need to read a figure carefully and match it to a claim's variable and direction. It's the textual-evidence skill with numbers attached.

#The method (read the figure BEFORE the choices)

Step 1 — Read the title, axes, and UNITS first. Before anything else: What is being measured? What are the two axes? What are the units (percent? thousands? grams per liter? years)? Which direction is "more"? Students misread quantitative items mostly because they never actually read the axis labels. Spend five seconds here and you'll skip most traps.

Step 2 — Read the legend/keys. If bars are grouped or lines are colored, know which is which (e.g., "solid line = Species A, dashed = Species B").

Step 3 — Pin the claim's variable and DIRECTION. Just like textual: what quantity, and which way (rising? higher than? lower than? no change?).

Step 4 — Find that variable in the figure and read the actual numbers. Point to the specific bar/point/cell. Note the value.

Step 5 — Test each choice for THREE things:

Is it accurate? (Does the figure actually say this? Wrong numbers = out.)
Is it relevant to the claim? (Right variable?)
Does it go the right direction to do the job (support/complete)?

A choice can be true to the graph but irrelevant to the claim, or relevant but misread the graph. You need all three boxes checked.

#Traps specific to quantitative evidence

Accurate but irrelevant. The choice correctly describes a number in the figure that has nothing to do with the claim. (Very common — a true statement that doesn't support the point.)
Misread value / wrong number. Sounds right but the number is wrong, or reads the wrong bar/row.
Direction reversed. Correct variable, but says "increased" when the figure shows a decrease, or "greater than" when it's less than.
Comparison the figure doesn't license. Compares two things the graph doesn't actually put side by side, or claims a trend from a single data point.
Confuses the two axes / two series. Reads Species A's value as Species B's, or reads the x-value as if it were the y-value.

Golden test for quantitative: "Does this choice (1) match the numbers in the figure, (2) concern the claim's variable, and (3) point the direction the task needs?" Three yeses or it's out.

#Worked Example 1 (bar-graph in words)

Passage: A student studied how four different soil types affected the average height of bean seedlings after 30 days. She concluded that clay soil produced the shortest seedlings.

Figure (described): A bar graph titled "Average Seedling Height by Soil Type (cm, after 30 days)." Four bars:

Sandy: 18 cm

Loam: 24 cm

Silt: 21 cm

Clay: 12 cm

Which choice best uses data from the graph to support the student's conclusion?

A) Seedlings in loam soil reached an average height of 24 cm. B) Seedlings in clay soil reached an average height of 12 cm, the shortest of the four soil types. C) Seedlings in sandy soil were shorter than those in silt soil. D) Seedlings in clay soil were taller than those in sandy soil.

Pin the claim. Conclusion: clay produced the shortest seedlings. Variable: seedling height; the claim is specifically about clay being the minimum. Prediction: I want the choice that states clay's value and notes it's the lowest.

A — Accurate but irrelevant. 24 cm for loam is a true number, but it's about loam, not about clay being shortest. Doesn't support this claim. Out.
B — Correct. Clay = 12 cm, and "the shortest of the four" is exactly the conclusion. Accurate ✓, relevant ✓, right direction ✓. In.
C — Accurate but irrelevant. Sandy (18) < silt (21) is true, but says nothing about clay being shortest. Out.
D — Reversed / wrong. Clay (12) is shorter than sandy (18), not taller. Misreads the direction and undercuts the claim. Out.

Trap contrast: A and C are both true statements about the graph — that's the whole trick of quantitative evidence. Truth isn't enough; the number has to serve the specific claim. Only B ties clay's value to the "shortest" conclusion. Always ask "does this number support the claim," not just "is this number right." B.

#Worked Example 2 (line-graph trend, "complete the text")

Passage: A city tracked the number of registered electric vehicles (EVs) from 2018 to 2023. A researcher noted that EV adoption accelerated over the period, and wrote that the yearly increase in registrations grew larger toward the end of the period. To support this, she pointed out that ______

Figure (described): Line graph, "Registered EVs in the City (thousands)," by year:

2018: 5

2019: 7

2020: 10

2021: 15

2022: 23

2023: 34

Which choice most effectively uses data from the graph to complete the text?

A) registrations rose from 5,000 in 2018 to 7,000 in 2019, an increase of 2,000, while from 2022 to 2023 they rose from 23,000 to 34,000, an increase of 11,000. B) registrations were higher in 2023 than in 2018. C) registrations increased every year from 2018 to 2023. D) registrations reached 34,000 in 2023.

Pin the claim. Claim: the yearly increase grew larger over time (i.e., the increments got bigger — acceleration). Variable: the year-over-year change, and the direction is increasing increments. Prediction: I need a choice that compares an early year-gap to a late year-gap and shows the late gap is bigger.

A — Correct. Early increment (2018→2019) = 2,000; late increment (2022→2023) = 11,000. Bigger yearly jumps later = acceleration. This is the only choice that speaks to increments, and the numbers are accurate. In.
B — Too weak / wrong variable. "Higher in 2023 than 2018" only shows growth, not accelerating growth. Any increasing series satisfies it. Doesn't support "increases grew larger." Out.
C — Too weak / wrong variable. "Increased every year" also shows only steady growth, not that the increases got bigger. Out.
D — Accurate but irrelevant. A single endpoint value says nothing about the pace of increase. Out.

Trap contrast: B, C, and D are all true to the graph. But the claim is subtle — it's about the change in the change (acceleration), not mere growth. Only A compares increments. When a claim is about a trend "speeding up" or "slowing down," the answer must compare the size of the gaps between points, not just the values. This is exactly the kind of precise distinction a math brain should dominate — read the claim as carefully as you'd read an inequality. A.

#Worked Example 3 (two-series table, direction trap)

Passage: Researchers compared two fertilizers, X and Y, on tomato yield across three field trials. They concluded that Fertilizer Y consistently outperformed Fertilizer X.

Figure (described): Table, "Tomato Yield (kg per plot)":

Trial Fertilizer X Fertilizer Y
Trial 1 30 34
Trial 2 28 31
Trial 3 33 37

Which choice best describes data from the table that support the researchers' conclusion?

A) In Trial 2, Fertilizer X produced 28 kg per plot. B) In each of the three trials, Fertilizer Y produced a higher yield than Fertilizer X. C) Fertilizer Y produced its highest yield in Trial 3. D) In Trial 1, Fertilizer X and Fertilizer Y produced similar yields.

Trial	Fertilizer X	Fertilizer Y
Trial 1	30	34
Trial 2	28	31
Trial 3	33	37

Pin the claim. Conclusion: Y consistently outperformed X. Key word: consistently = in every trial. Variable: Y vs. X yield, direction Y > X, across all three. Prediction: the choice that states Y > X in each trial.

A — Accurate but irrelevant. X's Trial-2 value alone doesn't compare Y to X. Out.
B — Correct. 34>30, 31>28, 37>33 — Y beats X in all three trials, which is exactly "consistently outperformed." Accurate ✓, relevant ✓, right direction ✓, and it covers all trials (the "consistently" part). In.
C — Accurate but irrelevant. Y's own maximum trial doesn't compare Y to X. Out.
D — Wrong / undercuts. 30 vs. 34 is a 4 kg gap; calling them "similar" both misreads and works against the claim. Out.

Trap contrast: The word consistently is load-bearing. A choice that showed Y beating X in one trial (or Y's own high point, like C) wouldn't establish consistency. B is the only one that checks every row. When a claim says "consistently," "always," or "in every case," the supporting data must cover all the cases, not just one. B.

#Quantitative Evidence — Mini-Practice (6 items)

P1. Passage: A student found that a plant flowered earlier as spring temperatures rose, concluding warmer springs led to earlier flowering. Figure: Table of average spring temperature (°C) and first-flower day (day of year): {8°C, day 110}, {10°C, day 102}, {12°C, day 95}, {14°C, day 88}. Which best supports the conclusion?

A) At 8°C, the plant first flowered on day 110.
B) As spring temperature rose from 8°C to 14°C, the first-flower day fell steadily from day 110 to day 88.
C) The plant flowered on day 95 when the temperature was 12°C.
D) The plant flowered every year.

P2. Passage: A shop owner claimed online sales made up a larger share of revenue each year from 2020 to 2023. Figure: Online share of revenue: 2020: 20%, 2021: 20%, 2022: 25%, 2023: 40%. Which best completes a supporting statement?

A) Online share was 20% in 2020 and 40% in 2023, rising in the later years.
B) Online share was 20% in 2021.
C) In-store sales were always larger than online sales.
D) Online share reached its lowest point in 2020.

P3. Passage: Researchers concluded that Species A tolerates salt better than Species B, because A survived at higher salinity. Figure: Survival rate at 30 g/L salinity — Species A: 80%, Species B: 45%. At 10 g/L — Species A: 95%, Species B: 92%. Which best supports the conclusion?

A) At 10 g/L, both species had survival rates above 90%.
B) At 30 g/L salinity, Species A's survival rate (80%) was much higher than Species B's (45%).
C) Species B survived at 45% at 30 g/L.
D) At 10 g/L, Species A survived at 95%.

P4. Passage: A student concluded that a metal's electrical resistance increased with temperature. Figure: Resistance (ohms) — 20°C: 4.0, 40°C: 4.6, 60°C: 5.3, 80°C: 6.1. Which best supports the conclusion?

A) At 20°C the resistance was 4.0 ohms.
B) As temperature rose from 20°C to 80°C, resistance rose from 4.0 to 6.1 ohms.
C) The metal conducts electricity.
D) Resistance was measured at four temperatures.

P5. Passage: A researcher claimed that Region P recycled a greater share of its waste than Region Q in every year from 2019 to 2021. Figure: Recycling rate — 2019: P 40%, Q 35%; 2020: P 42%, Q 38%; 2021: P 45%, Q 41%. Which best supports the claim?

A) In 2021, Region P recycled 45% of its waste.
B) In each year from 2019 to 2021, Region P's recycling rate exceeded Region Q's.
C) Region Q's rate rose from 35% to 41%.
D) In 2019, Region P recycled 40%.

P6. Passage: A student concluded that a battery's capacity declined faster after 300 charge cycles than before. Figure: Capacity (% of original) — 0 cycles: 100, 100 cycles: 97, 200 cycles: 94, 300 cycles: 91, 400 cycles: 84, 500 cycles: 77. Which best supports the conclusion?

A) From 0 to 300 cycles capacity fell by 3% each 100 cycles, but from 300 to 500 cycles it fell by about 7% each 100 cycles.
B) Capacity was 100% at 0 cycles.
C) Capacity declined over the battery's life.
D) Capacity reached 77% at 500 cycles.

#Quantitative Evidence — Answer Key

P1 — B. Shows the full inverse trend (temp up, flower-day down). A, C = accurate but single points, no trend. D = irrelevant.
P2 — A. Captures the rise (and honestly notes it happened "in the later years," matching flat-then-rising data). B, D = accurate but irrelevant single points. C = wrong variable (in-store vs. online, not share over time).
P3 — B. The high-salinity comparison is the claim. A, D = irrelevant (low salinity, both fine). C = one value, no comparison.
P4 — B. States the full rising trend. A = single point. C, D = irrelevant.
P5 — B. Covers "every year" with P>Q each time. A, D = single values, no comparison. C = only Q's trend, wrong focus.
P6 — A. Compares the rate of decline before vs. after 300 cycles (3% vs. 7% per 100) = "faster after." B, D = single points. C = only shows decline, not faster decline.

#PART 4 · THE OTHER READING TYPES (BRIEF)

These are generally more forgiving than Inferences and Command of Evidence, but the same golden rule applies: the answer is provable from the text.

#Central Ideas & Details

Method. For a main idea question, ask "what is the ONE point the whole passage builds toward?" The answer covers the whole passage, not one sentence — reject choices that are true but only about a detail. For a detail question ("According to the text…"), the answer is stated explicitly; go find the exact sentence and match it. No inference required — if you're reasoning hard, you're overthinking a detail question.

Example. Passage: "Although honeybees are the most famous pollinators, most of the world's wild plants are pollinated by a huge variety of lesser-known insects — flies, beetles, moths, and native bees — whose combined contribution far exceeds that of honeybees." Main idea? → "Lesser-known insects, collectively, pollinate more of the world's wild plants than honeybees do." (A choice saying only "Honeybees are famous pollinators" is a true detail, not the main idea — it ignores the "Although… far exceeds" turn.)

#Words in Context

Method. Cover the word. Read the sentence and predict your OWN plain word from context. Then pick the choice closest to your word. Ignore the word's most common meaning if context points elsewhere — these test the sentence's meaning, not dictionary popularity. Logic words and the surrounding contrast usually hand you the meaning.

Example. "The critic's praise was so ______ that readers couldn't tell whether she admired the novel or merely tolerated it; her review carefully avoided any strong statement." Blank ≈ "vague/noncommittal" (from "couldn't tell" + "avoided any strong statement"). Choose the option meaning noncommittal — not "generous" or "harsh," which the context rules out.

#Text Structure & Purpose

Method. For purpose, ask "why did the author write this — what are they trying to DO?" (describe a problem? challenge a common belief? give an example of a principle?). For structure, describe the shape ("presents a belief, then offers evidence against it"). Watch the logic words — they reveal the structure. Answer choices are often abstract ("To illustrate a limitation of a method"); match the function, not the topic.

Example. Passage: "Many assume twins raised apart prove genetics rules personality. But such twins often grow up in similar types of households, making it hard to separate genes from environment." Purpose? → "To point out a difficulty in interpreting a common type of evidence." (Not "to prove genetics don't matter" — the passage only complicates the interpretation.)

#Cross-Text Connections

Method. Two short passages by different authors on one topic. First, nail down each author's position separately in one phrase (agree? disagree? qualified?). Then determine the relationship (Text 2 supports / challenges / complicates / offers a different cause than Text 1). The stem often asks "how would the author of Text 2 respond to Text 1?" — answer using only Text 2's stated view applied to Text 1's specific claim.

Example. Text 1: "The city's falling crime rate proves the new policing strategy worked." Text 2: "Crime fell in the same years in dozens of cities that used no such strategy, suggesting a broader national trend." How would Text 2's author respond to Text 1? → "By noting that the drop may reflect a nationwide trend rather than the local strategy." (Text 2 offers an alternative cause — the exact confound logic from Part 1.)

#PART 5 · UNIVERSAL READING CHECKLIST + STUDY ORDER

#The universal per-question routine (run every time)

Read the stem first. Identify the type and the task verb (support / weaken / infer / main idea / purpose / meaning). Underline the verb.
Read the whole short passage once, carefully. No skimming — it's short by design.
Find the point and mark logic words (however, therefore, because, for example). The turn after "however" and the payoff after "therefore" are answer real estate.
For figures: read the title, axes, units, and legend before the choices.
Predict the answer in your own words before reading A–D. Non-negotiable. This is your anti-trap shield.
Eliminate three choices with text evidence. For each survivor, point to the exact words (or the exact number in the figure) that prove it. If you can't point, it's not it.
Name the trap in your eliminations: overreach, outside knowledge, half-right, reversed, off-topic. If your "answer" commits one of these, drop it.
Confirm with the golden rule: the winner is provable from the text alone; it needs no assumption, no added fact, no outside knowledge.
Watch the clock, but never skip prediction. Skipping step 5 is the #1 cause of inconsistency.

#Print-and-tape decision reminders

Uncertainty ≠ proof of the opposite. ("We can't confirm X" does not mean "X is false.")
Two things changing together ≠ one caused the other. (Confounds kill causal claims.)
True in real life ≠ supported by this passage.
Right topic, wrong direction = trap. Check the direction on every evidence question.
"Consistently / always / every" claims need evidence covering all cases.
For "illustrate," match the technique/function, not the prettiest wording.
Extreme words (always, never, only, proves, entirely, impossible) are usually wrong on Reading — the SAT lives in "suggests," "may," "tends to."

#The order to LEARN and DRILL these

Learn in the order that builds the core skill (provability) first, then layer on the specialized types:

Words in Context — quick confidence, teaches you to use context tightly. (A few days.)
Central Ideas & Details — trains "point of the passage" + literal matching. (A few days.)
Text Structure & Purpose — trains reading for function and logic words. (A few days.)
Command of Evidence — Textual — your first big battle; drill claim-pinning and task verbs until automatic. (1–2 weeks.)
Command of Evidence — Quantitative — same skill + figure-reading discipline; drill "read axes first." (1 week, alongside #4.)
Cross-Text Connections — combines the above (two positions + a relationship). (A week.)
Inferences — hardest, so drill it last and longest, once your provability instinct is solid. Do these in volume. (2–3 weeks, ongoing.)

Note the two hardest (Inferences and Command of Evidence) are trained near the end and get the most reps — but only after the "provable from the text" habit is second nature, because that habit is what makes all of them click.

#Practice discipline

Volume with review beats speed. After every practice set, keep an error log: for each miss, write the trap name (overreach? confound? direction? outside knowledge?) and the exact words that prove the right answer. Patterns in your log tell you your personal leak.
Slow down until you're accurate, then speed up. Consistency first; pace follows automatically once the routine is grooved.
Redo missed questions a week later with the routine, from scratch.

#Where to get REAL practice

This guide teaches the method with original examples. For authentic questions that exactly match the live test, use the official College Board Question Bank (searchable on College Board's site) and the practice tests inside the Bluebook app. Those are the gold standard for difficulty, phrasing, and figure style. Drill the method here; prove it on official items there. Nothing else — no third-party imitation — matches the real thing as closely, so make the official Question Bank your primary practice source and treat everything else as supplementary.

Master one habit above all: before you look at the answer choices, say the answer in your own words — then make the passage prove it. Do that every time and Reading stops being a coin flip and starts being a proof you can win.

MERIDIAN · CODEX · READING FROM FIRST PRINCIPLES | v1.0

Meridian Classic · progress saved in this browser · sign in to sync across devices

Up next

Command of Evidence — Textual + Quantitative

Match the evidence to the exact claim — text or data

5 min