AP Statistics exam guide 2027
The AP Statistics course was revised for the 2026–27 school year. This guide follows the new five-unit course framework and the new exam, with a free lesson for every topic, how answers are scored and a 10-week revision plan.
The exam at a glance
The May 2027 exam is fully digital in the Bluebook app, lasts 3 hours, and allows a graphing calculator. Reference sheets with formulas and tables are provided.
| Section | Questions | Time | Weight |
|---|---|---|---|
| Section I: Multiple choice | 42 questions | 1 hour 30 minutes | 50% of your score |
| Section II: Free response | 4 questions | 1 hour 30 minutes | 50% of your score |
The four free-response questions are: Question 1, a multi-focus question on formulating questions and collecting data; Question 2, a multi-focus question on analysing data and interpreting results; Question 3, an inference question (a confidence interval or a significance test); Question 4, a multi-focus question that pulls several skills together. The separate investigative task from the old exam has gone.
What changed in 2026The course went from nine units to five. The geometric distribution, combining random variables, the chi-square goodness-of-fit test and inference for the regression slope are no longer separate required topics. Our inference for regression step stays on the site for university students.
What is on the exam, unit by unit
Unit 1: Exploring One-Variable Data and Collecting Data 20–30% of the multiple-choice section
| Unit | What you must know | Learn it here |
|---|---|---|
| 1 | Variables, data, parameters and statistics | Statistical Thinking |
| 1 | Tables and graphs for one categorical variable | Describing Data with Graphs |
| 1 | Graphs for one quantitative variable and describing distributions (shape, centre, spread, outliers) | Describing Data with Graphs |
| 1 | Summary statistics, boxplots and comparing distributions | Describing Data with Numbers |
| 1 | Random sampling and sampling problems (bias) | Data and How It's Collected |
| 1 | Experimental design: random assignment, blocking, matched pairs, scope of inference | Experimental Design |
Unit 2: Probability, Random Variables and Probability Distributions 15–25%
| Unit | What you must know | Learn it here |
|---|---|---|
| 2 | Two-way tables and summary statistics for two categorical variables | Two-Way Tables and Simulation |
| 2 | Estimating probabilities with simulation | Two-Way Tables and Simulation |
| 2 | Probability rules, mutually exclusive and independent events, conditional probability | Probability |
| 2 | Random variables, mean and standard deviation, binomial distribution | Discrete Random Variables |
| 2 | Normal distribution | The Normal Distribution |
| 2 | Sampling distributions and the Central Limit Theorem | Sampling Distributions and the Central Limit Theorem |
Unit 3: Inference for Categorical Data: Proportions 15–25%
| Unit | What you must know | Learn it here |
|---|---|---|
| 3 | Estimators and sampling distributions for proportions | Sampling Distributions and the Central Limit Theorem, Confidence Intervals |
| 3 | Confidence interval for one proportion | Inference for One Proportion, Confidence Intervals |
| 3 | Significance test for one proportion and p-values | Inference for One Proportion, Hypothesis Testing: The Foundations |
| 3 | Type I and Type II errors, power | Errors and Power in Significance Tests |
| 3 | Interval and test for the difference of two proportions | Comparing Two Proportions |
| 3 | Chi-square tests for homogeneity and independence | Chi-Square Tests for Categorical Data |
Unit 4: Inference for Quantitative Data: Means 10–20%
| Unit | What you must know | Learn it here |
|---|---|---|
| 4 | Sampling distributions for sample means | Sampling Distributions and the Central Limit Theorem |
| 4 | t interval and t test for a mean or mean difference (paired data) | One-Sample t Procedures and Paired Data |
| 4 | Interval and test for the difference of two means | Comparing Two Groups |
Unit 5: Regression Analysis 10–20%
| Unit | What you must know | Learn it here |
|---|---|---|
| 5 | Scatterplots, correlation, least-squares regression and residuals | Exploring Relationships |
How free-response answers are scored
- Answers are scored E, P or I (essentially correct, partially correct, incorrect) on each part, then combined into a score for the question. Clear communication matters as much as the numbers.
- Always answer in context. “The mean is 4.2” earns less than “the mean commute time for these 30 students is 4.2 minutes”.
- Use the four-step inference structure every time: State the parameter and hypotheses (or confidence level), Plan by naming the procedure and checking random, 10% and normal/large-counts conditions, Do the calculation, and Conclude in context.
- Write the conclusion with the exact logic: “Because the p-value of 0.012 is less than α = 0.05, we reject H0. There is convincing evidence that the true proportion of …”
- Never say “accept H0” and never say the interval has a 95% chance of containing the parameter. The 95% describes the method.
- Calculator output alone is not enough. Name the test, give the test statistic and p-value, and label any numbers you write down.
10-week revision plan
| Week | Focus | Do this |
|---|---|---|
| 1 | Unit 1 | Describing distributions with SOCS (shape, outliers, centre, spread) in context. |
| 2 | Unit 1 | Sampling methods, bias and experimental design. Write two full experiment designs. |
| 3 | Unit 2 | Two-way tables, simulation and probability rules. |
| 4 | Unit 2 | Random variables, binomial and normal calculations. |
| 5 | Unit 2–3 | Sampling distributions, then one-proportion intervals. |
| 6 | Unit 3 | One- and two-proportion tests, errors and power. |
| 7 | Unit 3 | Chi-square tests for homogeneity and independence. |
| 8 | Unit 4 | t procedures, paired data and two-sample means. |
| 9 | Unit 5 and mixed | Regression, residuals and computer output. One timed multiple-choice set. |
| 10 | Free response | Timed free-response sets, marked against the scoring guidelines. |
Common mistakes that cost points
Common mistakeNot checking conditions, or checking them with the wrong values (use p₀ for the large-counts check in a test and p̂ in an interval).
Common mistakeInterpreting a p-value as the probability that H0 is true.
Common mistakeTreating paired data as two independent samples.
Common mistakeSaying correlation proves causation. Only a randomised experiment supports a cause-and-effect conclusion.
Common mistakeDescribing a distribution without context or without all four features.
Ready to practise at exam level?
Statistics Made Simple and its practice book give you 1,569 fully worked questions, datasets and a mock paper.
Further reading
Our earlier article on this exam, with more background: