The September 2026 SAT: the sitting students could not agree on

Harder, but not uniformly. Reports lean clearly towards difficulty rather than comfort, and the students who found it hard sound more distressed than August’s did. What separates September is that the reports contradict one another far more than usual.

Reading and Writing is where the agreement is. Module 2 drew steady complaints, vocabulary dominated them for the second sitting in a row, and several students who expected a comfortable section did not get one. Math is where the accounts split, with the back half of Module 2 drawing both the heaviest complaints and the claims that it was the easiest part of the day.

Two things were notably quiet. Timing barely came up, and inference questions drew almost no discussion at all, which is a real change from August, where the reasoning-step items were one of the features of the form. What students wanted to talk about instead was geometry and words.

More students asked which Module 2 they had been routed into than asked about any actual question on the exam. The method was always the same: list the topics you remember, compare them against someone else’s list, and conclude from the overlap whether you got the hard version.

The underlying concern is sound. Routing is real, it happens on the basis of Module 1, and the harder Module 2 is the one that opens the top of the scale. A student who lands there has already done well. If you spent the evening worried that a hard Module 2 was bad news, the opposite is true.

The method is what fails. A topic list is not evidence of routing, because topics vary between forms for reasons that have nothing to do with difficulty. Two students who both got the harder Module 2 can share almost no questions, and a student on the easier module can meet a topic that sounds advanced. That is precisely why September’s accounts disagree: students were matching topic lists across different forms and reading the mismatch as a verdict on the exam, or on themselves.

It also explains the second pattern in this sitting’s feedback, which is students hoping a question they missed was unscored. Both habits are the same move: an attempt to convert an uncertain outcome into a settled one before scores exist. Neither is answerable, and the hours spent on them are the most expensive hours in the whole process, because they are the only ones that could still have been spent studying.

Geometry and measurement dominated what students reported from Math, and the questions that hurt shared one structure. Each supplied a quantity that looks like the one the formula needs and is not.

A pyramid question gives a vertical height, but the lateral faces are triangles whose own height is the slant height, and the slant height has to be built from the vertical height and half the base edge before any surface area can be computed. A right-triangle question gives the two pieces the altitude cuts the hypotenuse into, and the altitude is neither their sum nor their average but their geometric mean. A mixture question gives concentrations as percentages by mass, and the equation needs grams of solute, so every percentage has to be converted before it can be added to anything.

In each case the arithmetic is easy and the conversion is the question. This is also why students reported that the calculator had less purchase than usual. A graphing tool will finish any of these problems in seconds once the relationship is written down, and it cannot tell you that the height you were handed is the wrong height.

The correction is a habit rather than a formula. Before substituting anything, say out loud what each given number measures and what the variable in the formula measures. When those two descriptions do not match, you have found the actual work.

The section leaned on solids and triangles, with circles, proportional reasoning and a two-way table question alongside them. The recurring shape was a figure or scenario described in words, with one measurement supplied that stands next to, but not in, the formula the question needs.

A smaller cluster worked the other way round, giving a property and asking which value makes it hold: the constant that leaves a quadratic with exactly one solution, the point that cannot satisfy an inequality whatever the constant. That shape has now appeared on every recent form and is worth treating as permanent.

Vocabulary carried this section as it carried August’s, but the traps were built differently. August’s wrong answers were near-synonyms, slightly too strong or too weak. September’s were frequently the word pointing the opposite way.

The clearest case set paradigmatic against idiosyncratic. One means serving as the typical example of a pattern, the other means peculiar to a single case, and they are close to opposites. A student who reads the passage for topic rather than for direction has no way to choose between them. The same structure appeared around elusive, where the competing option described the researchers rather than the thing that had escaped them, and around lament, where the alternatives split between expressing regret and simply listing.

Beyond vocabulary the section ran punctuation of a series with internal commas, verb form in a sentence that already had its main verb, and a literary excerpt where the question turned on what one sentence was doing rather than on what it said.

Original questions written around what this exam tested, one for each argument above. Nothing here is a real test question.

September is the clearest recent case of a sitting where the loudest question was unanswerable. Which module you were routed into cannot be recovered from a list of topics, and knowing it would not change a single mark. The students who spent that evening matching topic lists were doing careful work on the one part of the day that was already finished.

The part that is not finished is the setup habit. Every question that hurt on this form, in both sections, turned on reading a given correctly before using it: which height, which denominator, which direction the sentence points. That is learnable in an afternoon and it transfers to every form after this one, which is more than can be said for any conclusion about routing.