The AP Statistics Course and Exam Description effective for 2026–27 is the controlling framework for the May 2027 exam. It consolidates the course into five units, organizes performance around four statistical practices, removes several legacy topics, and changes the exam to 42 MCQs plus four 10-point FRQs completed fully in Bluebook.
What the 2026–27 CED actually organizes
Use the published ranges as emphasis ranges, not as a promise that a specific exam will contain an exact number of questions from each unit.
Search all 55 required 2026–27 CED topics
This navigator uses the official topic codes with concise paraphrased study labels. It also shows how many questions in the Salar Cafe 2027 MCQ bank are mapped to each topic, so missing coverage is visible instead of hidden.
The CED is a reasoning framework, not a formula inventory
Formulate Questions
Identify the investigative question, population, variables, parameters, and what evidence could answer it.
Collect Data
Reason about samples, randomization, experiments, bias, controls, and what conclusions the design permits.
Analyze Data
Use displays, summaries, probability models, simulation, regression and inference procedures appropriately.
Interpret Results
State conclusions in context, distinguish association from causation, and explain uncertainty and evidence.
How the CED becomes the May 2027 exam
| Exam component | 2027 structure | What the CED tells you to practice |
|---|---|---|
| Section I | 42 MCQs · 90 min · 50% | All five units and all four practices, including two 3-question shared-prompt sets. |
| FRQ 1 | 10 points | Formulate Questions + Collect Data. |
| FRQ 2 | 10 points | Analyze Data + Interpret Results. |
| FRQ 3 | 10 points | Statistical inference. |
| FRQ 4 | 10 points | Multi-content reasoning with practices 2–4. |
| Delivery | Fully digital Bluebook | Typed statistical explanations plus calculator/reference fluency. |
Do not let an older CED control 2027 preparation
Geometric distribution
Removed from required 2026–27 AP Statistics content.
χ² goodness-of-fit
Removed; do not confuse this with χ² homogeneity/independence, which remain relevant.
Regression-slope inference
The former Unit 9 inference for quantitative-data slopes is no longer required core.
Combining random variables
Removed as a required topic; use the current Unit 2 scope.
Departures from linearity
The old dedicated topic was removed; residual reasoning still matters for assessing models.
Use the official document without getting lost in it
- 1
Start with the five-unit overview
Use the unit names and weighting ranges to build your study map.
- 2
Attach a practice to every topic
Ask whether the task is mainly formulation, collection, analysis, interpretation, or a combination.
- 3
Separate course scope from exam mechanics
The CED defines what and how you learn; the exam page controls section counts, timing and delivery.
- 4
Mark removed topics in old resources
Keep useful foundational explanations but do not spend core review time on retired requirements.
Move from the CED to practice
How the five units feed the four statistical practices
The CED is easier to use when you separate content from practice. Units tell you what statistical ideas are in scope; practices tell you what the exam asks you to do with those ideas. A student can know a formula and still miss the question because the practice is wrong.
| Unit | Main 2027 content | High-value practice moves | Typical trap |
|---|---|---|---|
| 1 · 20–30% | One-variable data, sampling, experiments | Formulate, collect, describe, scope | Claiming causation/generalization the design cannot support |
| 2 · 15–25% | Two categorical variables, simulation, probability, random variables, binomial/Normal models, sampling distributions | Model, calculate, interpret probability | Using a model without checking its conditions |
| 3 · 15–25% | One/two-proportion inference plus χ² homogeneity/independence | Select method, test/interval, interpret | Confusing parameter, statistic, p-value, and confidence level |
| 4 · 10–20% | Inference for means | Distinguish one-sample, paired, two-sample | Treating paired data as independent groups |
| 5 · 10–20% | Regression analysis | Model, residuals, interpretation | Causal claims or regression-slope inference outside 2027 core |
