From May 2027, the AP Statistics free-response section is four questions in 90 minutes, worth 50 percent of your total score. Each question is worth 10 points, so each carries the same weight: about 12.5 percent of your exam score.
What separates the top scorers on AP Stats FRQs isn't computational speed. It's the ability to write answers that satisfy the College Board's strict requirements about context and interpretation. Students who can execute a two-sample t-test cleanly but write "the p-value is less than 0.05 so we reject" without stating the null, the alternative, the assumptions, or the conclusion in context of the study lose much of the available credit. The rubric expects correct communication as well as correct arithmetic.
What the FRQ section really is
Section II has four FRQs in 90 minutes, each worth 10 points. Question 1 focuses on formulating questions and collecting data. Question 2 focuses on analyzing data and interpreting results. Question 3 is an inference question (a hypothesis test or confidence interval). Question 4 combines data collection, analysis and interpretation across several content areas.
How the FRQs are scored
Each revised FRQ is scored out of 10 points using a question-specific scoring guideline, which replaces the old 0-to-4 holistic scale. As before, credit depends on both correct statistical work and clear communication in context, so a correct calculation with a vague or context-free conclusion will not earn full credit. Check the sample questions and scoring guidelines in the 2026–27 CED to see how points are allocated.
The inference framework: state, plan, do, conclude
For any FRQ involving a hypothesis test or confidence interval, AP scoring guidelines consistently look for four components, which many teachers call the State, Plan, Do, Conclude framework. Skipping any component is where communication points get lost.
State: name the parameter you're estimating or testing, and write the null and alternative hypotheses using both symbols and words in context. "H0: mu1 - mu2 = 0 (the mean cholesterol level is the same for patients given the new drug and patients given the standard drug); Ha: mu1 - mu2 is not equal to 0 (the mean cholesterol levels differ)."
Plan: name the procedure and check its assumptions. "Two-sample t-test for a difference of means. Random assignment: given in the problem. Sample size: n1 = 30 and n2 = 35 are both at least 30, so the sampling distribution of the difference in sample means is approximately Normal. Independence: samples are independent since assignment was random."
Do: compute the test statistic and p-value (or the confidence interval). Show the setup, not just the calculator output. "t = (x1 - x2) / sqrt(s1-squared / n1 + s2-squared / n2) = -2.34. df = 62 (from calculator). P-value = 0.023."
Conclude: state the conclusion in context of the study. "Since the p-value (0.023) is less than alpha (0.05), we reject the null hypothesis. There is convincing evidence that the mean cholesterol level differs between patients given the new drug and patients given the standard drug."
Leaving out any of these components (the parameter, the conditions, the mechanics, or the conclusion in context) usually costs credit, even when the arithmetic is right.
"In context" is the phrase you'll see over and over in AP Stats scoring guidelines. It means naming the variables and units from the specific problem, not the generic "the data" or "the population." A conclusion that could apply to any two-sample t-test hasn't been written in context.
Ready to boost your grades?
Join 1M+ students who have used Cognito to ace their exams.
Get started for free!
Strategy for the four questions
With four equally weighted questions in 90 minutes, a simple plan is about 20 to 22 minutes per question, leaving a few minutes to check your work. Read each question in full before you start writing, because parts often build on each other: misreading the setup in part (a) can carry through to later parts. Questions 1, 2 and 4 each combine several practices (designing a study, describing data, interpreting results), so expect to switch between calculation and explanation within a single question. On Question 3, write out the full inference structure: parameter and hypotheses, procedure and conditions, mechanics, and a conclusion in context.
Common mistakes that cost points
Writing conclusions not in context. "Reject the null" is not a conclusion. "There is convincing evidence that the mean recovery time differs between the two treatments" is a conclusion in context. The rubric will not award the conclusion point without the context.
Skipping assumption checks on inference procedures. Every hypothesis test and confidence interval has assumptions (randomness, independence, sample size or normality). Missing them costs a point even if the calculation is correct.
Confusing statistical significance with practical significance. "The result is statistically significant, so the difference is important" is wrong on the rubric. Statistical significance means the observed difference is unlikely to be due to chance; it says nothing about the size of the difference in real-world terms.
Just writing the calculator output. Name the procedure (or give the test statistic formula), report the test statistic and p-value, and label your values. Calculator syntax on its own, like "p = 0.023 from calculator", won't earn full credit.
Practice AP Statistics with our past papers
Our full library of released AP Statistics past papers with worked solutions shows the exact state-plan-do-conclude framework and in-context conclusions the rubric rewards.
