"The SAT is adaptive" is true and, on its own, misleading. It leads students to picture a test that gets harder every time they answer correctly, which is not what happens and produces some genuinely counterproductive tactics. Here is the actual mechanism and what follows from it.
The mechanism
College Board calls the model multistage adaptive testing, and it explicitly contrasts it with the question-by-question kind. The digital SAT uses a two-stage version.
Each section is split into two separately timed modules. Reading and Writing is 54 questions in two 32-minute modules; Math is 44 in two 35-minute modules.
- Module 1 is the routing module. It carries half the section's questions and deliberately spans the full difficulty range — easy, medium and hard mixed together. Its job is to measure you, not only to score you.
- The app then chooses module 2. Based on your performance so far, it serves a second module that is, on average, either more difficult or less difficult than the first.
- Module 2 completes the section. Which form you were routed into affects the score range reachable in that section.
The two sections route independently. Nothing about your Math routing affects Reading and Writing, or the reverse.
What this does not mean
The questions do not adapt within a module. Once a module starts, its questions are fixed. Answering three in a row correctly does not summon harder ones. This matters practically: you can move freely inside a module, skip, flag, change answers and revisit right up until that module's timer runs out. Students who believe the test is question-adaptive often answer strictly in order and never use the flag, which costs them.
It is not computer-adaptive in the GRE sense of a fresh difficulty decision after every item. Two stages, one decision point per section.
The consequence that matters: module 1 has leverage
Module 1 is half a section's questions, so on a naive count it is half the section. In practice it carries more weight than that, because it also decides which second module you see, and being routed into the lower-difficulty form caps what the section can reach.
For a student allocating limited preparation time, that is a real planning input. Accuracy on medium-difficulty questions early in a section is worth more than the ability to crack the very hardest items, because the hardest items only appear in quantity in a module you have to be routed into first.
Do not try to game the routing
A tactic surfaces periodically on forums: deliberately underperform in module 1 so that module 2 is easy, then score highly on it. This does not work. The easier second module carries a lower reachable ceiling, so the plan trades a higher possible score for a more comfortable half-hour. It is a way of choosing a lower score on purpose.
What it changes about pacing
Because each module has its own clock and nothing carries over, time management has to be planned per module rather than per section. A student who finishes module 1 with eight minutes to spare has not banked anything — those minutes disappear.
The practical rule that follows: in module 1, spend the spare time checking rather than sitting idle. It is the one place on the exam where re-reading your own work has compounding value, because it feeds a routing decision as well as a score.
What it changes about practice
Only an adaptive practice test reproduces the routing decision, so only an adaptive score is comparable to a real one. A non-adaptive paper practice test is perfectly good for drilling content, but its total should not be read as a predicted SAT score.
This is the main reason to take full-length practice tests in Bluebook rather than from a book: College Board's full-length practice tests are free, adaptive and run in the same application as the exam.
The short version
98 questions, four modules, two routing decisions. Nothing adapts inside a module, so use the flag and move freely. Module 1 accuracy has leverage beyond its question count. Do not sandbag it. And take your measurement tests in the adaptive format, because a non-adaptive total is not the same number.
More on the exam's structure in our Digital SAT guide, and on turning a practice-test result into a plan in the error analysis guide.