SAT Practice Exams · How Your Exam Is Graded
Accredify offers two kinds of SAT practice exam. This guide shows, side by side, how each one picks your Module 2 and works out your score.
The College Board Equivalent exam copies the real digital SAT. It counts how many Module 1 questions you got right and uses that count to send you to one of two versions of Module 2.
The Beyond College Board Granularity exam measures you more finely. It never uses a flat count of correct answers. Instead, your ability estimate (called theta, written θ) decides every step, and it chooses between five versions of Module 2.
Both exams have the same shape: Reading and Writing (27 questions per module) and Math (22 questions per module), each split into Module 1 and an adaptive Module 2. They differ in how Module 2 is chosen and how the final score is worked out.
Module 1 is graded as a raw count: how many questions you answered correctly. Reach the threshold and you get the Hard Module 2; fall short and you get the Easy / Standard Module 2.
The section is then scored once with Item Response Theory, but the route has already set a ceiling: the Easy / Standard route tops out at around 650–680.
Module 1 is graded as an ability estimate (θ). Every question's difficulty, how sharply it separates strong and weak candidates, and how guessable it is all count, not just whether you got it right.
That θ picks one of five Module 2 variations, each aimed at a narrower ability band. Module 2 is scored the same way, so θ decides every step.
| Stage | College Board Equivalent | Beyond College Board Granularity |
|---|---|---|
Grading Module 1 |
Flat count of correct answers out of all 27 (R&W) or 22 (Math). Every question is worth the same. |
θ estimate from the 3-parameter IRT model. A hard, sharply discriminating question moves θ more than an easy one. |
Routing signal |
One threshold: 15 of 27 (R&W) or 13 of 22 (Math). |
Four θ boundaries: −2, −1, +1, +2. |
Module 2 options |
2: Easy / Standard and Hard. |
5: Variation 1 (most accessible) to Variation 5 (most challenging). |
Final score |
One θ per section from both modules, turned into a 200–800 score. The Easy / Standard route is capped at about 670. |
A θ for each module, turned into 100–400 and added into a 200–800 section score. No route cap: the variation's difficulty is already built into θ. |
During the exam |
Nothing about routing. As on the real SAT, Module 2 is chosen behind the scenes. |
You see which of the five variations you were given. |
When scores appear |
After the whole exam. |
After the whole exam. |
College Board splits Module 1 results into two buckets. The granular exam splits the ability scale into five bands, so you are matched to a Module 2 much closer to your actual level. Both bars below are drawn to scale.
One question either side of 15 changes the route. Math works the same way with 13 of 22.
θ runs from −3 to +3. Each band leads to a Module 2 built for that ability range.
Both students below answer 16 of 27 Reading and Writing Module 1 questions correctly. To College Board's routing they are identical. Their answers tell a different story.
correct, easier question correct, harder question missed
Missed several easy questions but got most of the hard ones: careless on the basics, strong on the difficult material.
Answered nearly every easy question and few of the hard ones: secure on the basics, still building on the harder material.
With a flat count, both students sit the same Hard Module 2. With θ, Student B gets a Module 2 pitched at their level, where each answer tells us more about what they know.
The θ values are illustrative. The real estimate depends on each question's calibrated difficulty, discrimination and guessing values.
Under College Board routing, a student with 14 correct and a student with 15 correct are nearly the same ability. But one is sent to the Easy / Standard module and capped at around 670, while the other can still reach 800.
The granular design has no single cliff. A θ just either side of a band boundary leads to neighbouring variations of similar difficulty, and the score is never capped by the route.
| College Board Equivalent | Beyond College Board Granularity | |
|---|---|---|
Module 2 routing | R&W: 15+ of 27 → Hard | V1: θ < −2.0 |
Score formula | Section = 500 + 100θ (one θ from both modules) | Module = 100 + 50(θ + 3); section = Module 1 + Module 2 |
Route cap | Easy / Standard: 670 | None |
Shared by both | 3-parameter IRT model · maximum-likelihood estimate · θ from −3 to +3 · 2 unscored pretest questions per module · section 200–800, total 400–1600 | |