Last updated: October 2026 · By the AllSatPapers Content Team
September 2026 9D combines three kinds of precision: keeping a research conclusion within the evidence, separating the grammatical levels in a complicated sentence, and labeling the quantities in a mathematical model. Its extra blocks make it useful for comparison practice, provided the two occurrences of Module 2 are recorded separately.
The six examples deliberately extend beyond the opening vocabulary sequence. The moth study asks how new evidence changes what researchers know; the programming-language sentence uses semicolons inside a complex list; the rational function and mixture questions require an intermediate result to be interpreted before it becomes the final answer. These are different reasons an otherwise fluent student can lose a point.
Plan your next practice session: Explore the September 2026 Digital SAT papers and bundle options.
What the Current File Contains
The file contains 145 questions across 151 pages: 81 Reading and Writing questions and 64 Math questions, with six key pages. Reading blocks begin on pages 1, 29, and 57 and have keys on pages 28, 56, and 84. Math blocks begin on pages 85, 108, and 131, with keys on pages 107, 130, and 151.
The final Math Module 2 block contains 20 questions rather than 22. That is what the supplied file displays; it should not be silently padded to a standard length. The repeated Module 2 blocks and their adjacent keys allow clear page-based review, but their labels alone do not establish a student’s adaptive path or a verified administration.
The date and regional designation follow the supplied filename. They identify this archive file, rather than independently authenticate a test administration. All current-paper page references are PDF pages.
Reading the Sections as a Whole
Reading and Writing: expanding a finding is not the same as explaining a mechanism
The moth question on page 14 uses a broader survey to suggest that an ultrasonic defense is more prevalent than earlier observations showed. It does not establish that every moth lacking that behavior lacks all defenses. The bee question on page 32 similarly requires a controlled scope: a haplotype can be prevalent in some Southern states while other haplotypes are most frequent nationally.
These two examples are useful together because both turn on the population named in a claim. One compares the range of previously known and newly observed genera; the other compares regional and national distributions. A word such as although does not automatically make the regional finding abnormal, just as a larger survey does not prove a universal rule. Read the qualifier attached to each conclusion.
The programming list on page 71 then shifts from scientific scope to sentence hierarchy. Each language has a developer-and-date description containing commas, so the main list boundaries need semicolons. A rule learned only as “semicolon equals two sentences” would miss this legitimate list use. The file rewards recognizing the construction before applying the rule.
Math: the requested quantity determines when the solution is finished
The skating model on page 133 asks for an equation, so solving for the original group sizes is unnecessary. The rational function on page 147 asks for another zero, so the y-intercept is an intermediate route to the numerator’s constant term. The mixture on page 150 asks for salt mass contributed by one solution, so the solution mass obtained from the linear equation is only an intermediate result.
These questions are compact but expose different stopping errors. In the skating item, a student may overwork a model-construction task. In the rational item, a student may accept a denominator zero as an intercept. In the mixture item, a student may stop too early and report grams of solution. Labeling every result with its role and units makes the appropriate stopping point visible.
The final 20-question block should be reviewed as it exists. Do not compare its raw correct count directly with a 22-question block without noting the different denominator. More useful than a blanket difficulty label is a record of whether the error came from a percentage multiplier, a domain restriction, or a solute-versus-solution distinction.
Worked Questions: Reasoning, Traps, and Efficient Methods
1. A wider survey changes what was believed to be rare
2026 9D, Reading and Writing Module 1 Q14, PDF p.14
A survey finds defensive ultrasonic clicks in 52 moth genera, including several beyond the groups previously known to use them.
Answer: B, the defense may be more prevalent among large-bodied moths than researchers had known. The first reading key on p.28 gives B.
The passage begins with a defense observed in a limited set of moth groups, then broadens the investigation to 252 genera. Finding the behavior in 52 genera, including several geometer genera, expands the known distribution. The supported conclusion concerns the prevalence of this particular defense. It does not require most genera to use it, only more than the previous account recognized.
A treats absence of recorded ultrasound as absence of any defense against bats. That is an unjustified leap because moths could have other defenses. C asserts a diverse range of effective strategies, while the study described measured ultrasonic responses. D proposes other purposes for ultrasonic signals, but the experiment used bat-echolocation playback and supplies no evidence about unrelated signaling functions.
Keep three quantities separate: all moth defenses, defensive ultrasound, and the groups in which that ultrasound was previously known. The result updates the third without establishing sweeping conclusions about the first. This makes a useful example of how a new study can expand a classification rather than overturn an entire explanation. You can answer from the before-and-after scope of the evidence without remembering the genus count as a fact for future tests.
2. An overall pattern can coexist with a local concentration
2026 9D, Reading and Writing Module 2 Q4, PDF p.32
C1 and C2j are most frequent overall; A1v is ___ in certain Southern states.
Reasoned answer: “prevalent.” The first Module 2 key on p.56 gives A, “prevalent.”
The sentence contrasts a national pattern with a regional one. Although two haplotypes are most frequent in the population overall, another occurs commonly in particular Southern states. Prevalent means widespread or common in the relevant place, so it fits the contrast without implying that A1v must be most common everywhere.
Elusive suggests difficulty finding or identifying something; anomalous suggests an unusual deviation; conclusive describes evidence that settles a question. None supplies the straightforward regional-frequency relationship established by the sentence. Anomalous can be tempting because although signals contrast, but a contrast between national and regional distributions does not make the regional pattern abnormal.
The efficient prediction is “common there.” Always keep the qualifier attached to the claim: a variant may be uncommon nationally and common locally at the same time. This scope distinction also matters in statistics and data interpretation. The haplotype labels are background identifiers, not vocabulary to study. If a later version changes those labels or the state, the same national-versus-local comparison still governs the choice.
3. Use semicolons to separate internally punctuated list items
2026 9D, Reading and Writing Module 2 Q15, PDF p.71
COBOL, Harbour, and Clojure are each followed by a developer-and-date description.
Answer: D, the option separating the three described languages with semicolons. The second Module 2 key on p.84 gives D.
The sentence’s main structure is “The programming languages COBOL; Harbour; and Clojure are all routinely translated…” Each language comes with a nonessential descriptive phrase containing commas. Semicolons distinguish the larger list boundaries from those internal commas. The correct choice closes the first item after 1959, separates the second after 1999, and leaves Clojure followed by its own descriptive comma.
The wrong options mix boundary levels. One uses a semicolon for the first separation but a comma for the next; another detaches developed from the language it describes; another omits a comma needed to enclose a descriptive phrase. Looking only at the nearest year cannot resolve these differences. You must map all three items and their attached descriptions.
Use a three-column mental list: language, developer, year. Then restore the sentence around it. The semicolons here are list separators rather than boundaries between independent clauses, so the common simplified rule about semicolons joining complete sentences is not sufficient. This question rewards the same structural reading used for a list of books with descriptive clauses, even though the subject matter is programming.
4. Keep each group’s growth factor attached to that group
2026 9D, Math Module 2 Q3, PDF p.133
A team starts with 44 skaters. Advanced skaters increase 53%, intermediate skaters 19%, and the total 32%.
Answer: C, 1.53a+1.19b=44(1.32). The final math key on p.151 gives C.
The original groups contain a and b skaters, with a+b=44. Their new sizes are 1.53a and 1.19b because an increase includes the original 100% plus the stated percentage. Adding the new groups must equal the new total, 44×1.32. This produces the displayed equation directly from the definitions.
The distractors swap the advanced and intermediate growth factors, attach the overall percentage to one group, or use only the percentage increase as though it were the entire new group. These are labeling errors rather than difficult arithmetic. Write a short annotation above each term—new advanced, new intermediate, new total—before choosing an equation.
The wording says approximately, so do not force the resulting model into an exact integer roster unless a separate question asks for one. A weighted average of the two group growth rates must lie between them, which provides a reasonableness check on the overall rate. The task is to construct the model, not solve unnecessarily for the initial counts. Correct labels are the fastest route through a familiar percentage story.
5. Use an intercept to recover the numerator, then check the domain
2026 9D, Math Module 2 Q17, PDF p.147
v(x)=(x²+bx+c)/[(x+4)(x−17)] passes through(0,55/68) and(11,0).
Answer: A, the other possible zero is−5. The final math key on p.151 gives A; the valid numerator root is −5.
At x=0 the denominator is−68, so c/(−68)=55/68 and c=−55. At x=11 the denominator is nonzero, so a zero output requires the numerator to be zero. Thus 11 is one root of x²+bx−55. The product of its roots is−55, making the other root−5. Equivalently, the numerator is(x−11)(x+5), giving b=−6.
Now apply the rational function’s domain restriction. The denominator is zero at−4 and 17, so those values cannot be zeros of the function. In particular, the offered−4 is an undefined input, not an x-intercept. At x=−5 the denominator is nonzero, so the numerator’s root is valid.
You could solve for b explicitly by substituting 11, then factor the quadratic; the product-of-roots shortcut simply avoids an unnecessary intermediate calculation. Either way, keep numerator zeros and denominator exclusions separate. The graph’s y-intercept supplies the constant term, while the x-intercept supplies one root. The difficulty is combining those representations in the right order, not expanding the denominator into a longer expression.
6. Distinguish the mass of a solution from the mass of its salt
2026 9D, Math Module 2 Q20, PDF p.150
Mix 0.30% and 0.25% solutions to obtain 90g containing 0.24g sodium chloride. Find the salt contributed by the 0.30% solution.
Answer: 0.09 grams of sodium chloride. The final math key on p.151 gives 0.09, consistent with the salt-balance calculation.
Let x be the mass in grams of the 0.30% solution. The other solution has mass 90−x. Convert the percentages to decimal fractions before writing the salt balance: 0.003x+0.0025(90−x)=0.24. Solving gives x=30. That is the mass of the first solution, not yet the requested mass of sodium chloride.
Multiply once more by its concentration: 0.003×30=0.09 grams. The remaining 60 grams contribute 0.15 grams of salt; the two contributions add to 0.24, checking the model. The final distinction is easy to miss because solving the linear equation feels like the end of the problem.
Two errors deserve separate attention. Treating a percentage such as 0.10% as 0.10 makes the concentration one hundred times too large. Returning x gives grams of solution instead of grams of solute. Label both equations and the requested unit. A table with columns for solution mass, concentration, and salt mass is a useful alternative setup, but the conservation relationships remain the same: total mass and total salt must each balance.
Earlier Papers: Repeats, Retuned Questions, and Useful Analogues
The comparisons cover March and May 2026 as well as August, September, and November 2025. Some reuse a mathematical model closely; others show how a similar scientific setting can test a different inference.
Moth defenses: prevalence versus mimicry
Related research setting, different inference: 2026 9D, Reading and Writing Module 1 Q14, PDF p.14 ↔ 2026 5C, Reading and Writing Module 1 Q14, PDF p.14.
May 2026 5C describes palatable calpine moths whose ultrasonic signals resemble those of noxious dalcerid moths. The supported inference is protective mimicry: a palatable species benefits from resembling a defended one. The current survey instead expands the known distribution of ultrasonic defense across genera. An option about greater prevalence is appropriate here but does not answer the older mechanism question. Compare what the study changed—knowledge of where a behavior occurs versus an explanation of why a species benefits from it—before transferring a conclusion.
Regional prevalence remains compatible with a different national pattern
Retuned vocabulary context: 2026 9D, Reading and Writing Module 2 Q4, PDF p.32 ↔ 2025 8D, Reading and Writing Module 1 Q5, PDF p.3.
August 2025 8D uses haplotype A4 in Louisiana, while the current version uses A1v in its Southern-state example. Both contrast the most frequent national haplotypes with one common locally. Prevalent expresses that local frequency. The changed labels are identifiers rather than new reasoning requirements. The useful transfer is keeping the geographical qualifier attached: national rarity would not prevent regional prevalence. Elusive and anomalous add meanings the contrast alone does not establish.
Complex lists have an outer and inner punctuation level
Skill analogue: 2026 9D, Reading and Writing Module 2 Q15, PDF p.71 ↔ 2025 9G, Reading and Writing Module 1 Q20, PDF p.13.
The September 2025 planetary-discovery item introduces a list of exoplanets with distance descriptions. A colon follows the complete introductory clause, and semicolons organize the descriptive items. The current programming sentence places descriptive developer-and-date material inside each language entry. Both require distinguishing punctuation within an item from punctuation between items. The exact blank differs, so the older colon is not the answer here; its list structure helps explain why the current semicolons are needed.
Three rates must keep their original labels
Retuned percentage model: 2026 9D, Math Module 2 Q3, PDF p.133 ↔ 2025 8E, Math Module 2 Q4, PDF p.60.
August 2025 8E starts with 48 skaters, raises advanced and intermediate counts by 27% and 54%, and raises the total by 42%. The corresponding model is 1.27a+1.54b=48(1.42). The current totals and rates produce 1.53a+1.19b=44(1.32). The model’s structure transfers exactly, but every multiplier must be rebuilt from the current group label. A familiar story is particularly dangerous when an option swaps two plausible percentages.
The intercept recovers a constant before a root relationship is used
Retuned rational function: 2026 9D, Math Module 2 Q17, PDF p.147 ↔ 2025 11E, Math Module 2 Q17, PDF p.56.
The November 2025 rational function has denominator (x+5)(x−19), y-intercept 66/95, and a zero at 11. Substitution gives c=−66, so the other numerator root is−6. In 9D the denominator at zero is−68 and the y-intercept is 55/68, producing c=−55 and the other root−5. The older excluded input−5 becomes a valid zero in the current question because the denominator has changed. That is why domain restrictions must be recomputed rather than remembered.
The older answer is solution mass; the current answer is salt mass
Retuned model with a changed requested output: 2026 9D, Math Module 2 Q20, PDF p.150 ↔ 2025 11C, Math Module 2 Q16, PDF p.69.
November 2025 11C mixes 0.30% and 0.15% solutions into 80 grams containing 0.21 grams of salt and asks for the mass of the 0.30% solution. That answer is 60 grams. The current concentrations and totals give 30 grams of the 0.30% solution, but the prompt asks for its sodium chloride contribution:0.003×30=0.09 grams. The conservation equations transfer, while the final requested unit changes. This is more than replacing numbers; it requires deciding whether the intermediate x already answers the question.
Five Ways to Use This Paper Well
- Use the adjacent key after completing a block, and record the final math block as 20 questions.
- In a research inference, state whether the new evidence changes prevalence, mechanism, or reliability.
- For complex lists, identify each complete item and its internal commas before selecting punctuation.
- In percentage equations, write the group name above each multiplier so plausible rates cannot be swapped.
- For rational functions and mixtures, label the intermediate result and ask whether it is the requested final quantity.
9D is especially useful for students whose mistakes occur after they recognize the broad topic. Recognizing “moths,” “percent increase,” or “quadratic numerator” is only the beginning. The point comes from maintaining the right population, grammatical level, domain, or unit through the final step.
Every Digital SAT Paper in Our Archive
Our archived files are grouped below by their supplied labels; regional labels and dates describe the archive, rather than independent authentication of an administration.
2023
- March 2023 (3A) · May 2023 (5A) · June 2023 (6A, Easy & Hard) · October 2023 (10A) · November 2023 (11A) · December 2023 (12A–12D)
2024
- March 2024 (3A–3D) · May 2024 (5A–5D) · June 2024 (6A–6E) · August 2024 (8A–8I) · October 2024 (10A–10G) · November 2024 (11A–11F) · December 2024 (12A–12H)
2025
- March 2025 (3A–3H) · May 2025 (5A–5F) · June 2025 (6A–6E) · August 2025 (8A–8F + 8G Question Bank) · September 2025 (9A–9I + 9J Question Bank) · October 2025 (10A–10H, 10I/10J School Day, 10K Question Bank) · November 2025 (11A–11G) · December 2025 (12A Math only, 12B English only, 12C, 12D, 12E, 12F Question Bank)
2026
- March 2026 (3A Question Bank, 3B, 3C US, 3D US English only) · May 2026 (5A, 5B, 5C–5H English only) · June 2026 (6A–6D, 6E Math only, 6F/6G English only, 6H Question Bank, 6I–6K) · August 2026 (8A–8D International, 8E collection labeled “4Sets”, 8F/8G US, 8H International Question Bank) · September 2026 (9A–9C International, 9D–9H US, 9I International Question Bank) · October 2026 (10A–10D International, 10E International English only, 10F–10I US, 10J International Question Bank, 10K US Question Bank; School Day files labeled October 8, October 13 sets 1/2, October 14, October 17 sets 1/2).
Build a practice sequence around the skills you need. Explore September 2026 SAT papers and bundle options, then revisit the matched earlier questions with the current numbers and wording in view.
Keywords: Digital SAT September 2026, SAT 9D, SAT question analysis, SAT past papers, allsatpapers.com
Last updated: October 2026. All test content referenced under fair use for educational analysis. SAT® is a registered trademark of the College Board, which was not involved in the production of, and does not endorse, this analysis or any products offered on this site.