How AP US History is actually scored
Your AP US History score isn't a percentage of questions right. It's a weighted composite of four separately graded sections, converted to a 1-5 scale by cutoffs the College Board resets every June.
The exam runs three hours and fifteen minutes across four parts, and each part carries a fixed share of your score.
| Section | Raw max | Weight | Per point |
|---|---|---|---|
| Multiple choice | 55 | 40% | 0.73% |
| Short answer (3 × 3) | 9 | 20% | 2.22% |
| Document-based question | 7 | 25% | 3.57% |
| Long essay question | 6 | 15% | 2.50% |
Each raw section score gets scaled to its weight, then the four are summed. That composite is traditionally expressed out of 130 points, though the scale itself is arbitrary - what matters is the percentage. This calculator shows you both.
Why the APUSH curve moves
APUSH is not curved against other students. The College Board uses a process called equating, where a set of repeated questions carried across years lets them measure whether this year's cohort is stronger and whether this year's exam ran harder. The Chief Reader then sets composite cutoffs after the June AP Reading, once every essay in the country has been graded.
That process produced a real shift recently. In 2023, 29% of APUSH students scored a 1 and fewer than half passed. In 2025, only 8% scored a 1, 73% passed, and 36% walked away with a 4. Some of that is genuine improvement in student performance. Some of it is the College Board's broader re-standardization of AP scoring against college outcomes data.
The practical consequence is that any calculator still running pre-2024 thresholds will under-predict you, sometimes by a full point. That's why this one gives you both bands and flags the cases where your result depends on which standard you believe.
Grading your own DBQ honestly
The biggest source of error in a predicted score isn't the curve. It's students grading their own essays generously. The DBQ rubric isn't holistic - it's six discrete checkboxes, and readers award each one independently. You either did the thing or you didn't.
The two points most often left behind
Sourcing asks you to explain, for at least two separate documents, how the point of view, purpose, historical situation, or intended audience matters to your argument. Naming the author isn't sourcing. Writing "this source is biased" isn't sourcing. You have to explain why it changes the way the document should be read.
Complexity used to be the rarest point on the exam, and it's still the one most often missed - but it got considerably more reachable in 2024. You can earn it with genuine nuance, analyzing multiple variables or confronting evidence that cuts against your own thesis. You can also earn it by using all seven documents effectively, or by sourcing four of them rather than the two the sourcing point requires. That last route means a strong sourcing habit can quietly pay twice.
If you're unsure whether you earned a point, mark it zero. A prediction built on optimistic self-grading is worse than no prediction at all.
Check which rubric you're working from
The College Board revised the AP History rubrics in 2024, and a lot of practice material still floating around uses the old one. Two rows changed. The second evidence point dropped from six documents to four, and the sourcing point dropped from three documents to two. If a workbook or teacher handout tells you that you need six documents, it predates the change. The rubric above is the current one.
APUSH scoring questions
How many multiple-choice questions do I need for a 5?
There's no fixed number, because the MCQ section is only 40% of your score. A student with 40 of 55 correct and strong essays clears a 5 more comfortably than a student with 50 correct and weak ones. Use the gap readout above - it converts the distance to your next score into the specific number of MCQs, SAQ points, DBQ points, or LEQ points that would close it.
Is there a guessing penalty?
No. The quarter-point deduction for wrong answers was removed in 2011. Answer every question. A blank and a wrong answer are worth exactly the same, so a random guess is free expected value.
Do I have to answer all four short-answer questions?
No. Questions 1 and 2 are required, then you choose between Question 3 and Question 4. You answer three in total, each worth 3 points, for 9 possible.
Does the DBQ have to use every document?
No. You need to accurately describe three documents for the first evidence point, and use four to support your argument for the second. Working all seven in is one of the ways to earn the complexity point, so it's worth attempting if the documents genuinely fit - but stretching to include one you don't really understand usually costs more than it gains.
How accurate is this prediction?
The section weights are exact - those come straight from the College Board. The uncertainty sits in two places. First the cutoffs, which are estimates. Second your own essay self-scoring, which is almost always the larger error. If the calculator puts you within a couple of points of a boundary, treat both outcomes as live.
Can I still get a 5 if I bombed the LEQ?
Yes. The LEQ is only 15%, the lightest section on the exam. A zero there costs you 15 percentage points, which is survivable with a strong MCQ and DBQ. A zero on the DBQ is much harder to come back from.