Interpret the result
A correct numerical value still needs a conclusion in context. The reader should be able to tell what the result says about the population, parameter, or scenario in the question.
Oxbridge Math Tutors AP Statistics · Free-response resource
A calculation is not the whole answer.
A plain-language guide to the reasoning, interpretation, assumptions, and written conclusions that matter in AP Statistics free-response questions.
Check the conditions
Strong responses make the relevant assumptions visible. The method is not just a button to press; it depends on whether the conditions support the inference.
Write for the mark
Free-response work rewards a sequence of reasoning: identify the method, show the evidence, calculate accurately, and explain what the result means.
How scoring works
Historical scoring versus the May 2027 exam
Do not mix the two. The E/P/I idea below describes 2026-and-earlier papers only.
Historical
2026 and earlier: historical
Each question had a guideline judging its parts as Essentially correct, Partially correct or Incorrect. The final 0 to 4 score was derived from those judgements for that question. It was never a count of keywords.
Verified: AP Central
May 2027: four FRQs, 10 points each
The exam has four typed FRQs in Bluebook, each worth 10 points. We do not present the older E/P/I scheme as the 2027 mechanism. Consult the latest official sample rubric and scoring guidance on AP Central for this format.
No universal buzzwords
Words like "normal" or "independent" earn nothing alone. The reader looks for reasoning tied to the information in the question.
Practice habit
Compare your response with the official guideline
Write first, then read the guideline
- Cover the guideline and write the full response.
- Mark each part against that question’s official guideline (for 2026 and earlier papers, E, P or I; for 2027 practice, the relevant sample rubric).
- Rewrite any part that did not meet the relevant question’s scoring requirements, and note why.
Keep an error log
- Wrong procedure chosen
- Condition missing or unchecked
- Conclusion not in context
- Claim stronger than the design allows
Worked educational example
A two-sample comparison, written for the reader
A school randomly assigns 40 students to practise recall with spaced sessions and 40 to practise in one block. After two weeks the spaced group has a higher mean quiz score. Does the evidence support a conclusion about the method?
- ContextCompare the mean quiz scores for students like these under spaced versus blocked practice.
- DesignRandom assignment to treatments makes this an experiment, so a causal claim is available if the difference is larger than chance variation would explain.
- Procedure and conditionsA two-sample t procedure is a candidate. Check the relevant conditions: random assignment, independence between groups, and that the score distributions give no strong skew or outliers at these sample sizes.
- ConclusionIf the test gives strong evidence, say that the data suggest spaced practice raises the mean score for students similar to these. Do not extend the claim to all schools: the 80 students were not a random sample of a wider population.
Scoring note: this is a reasoning example, not a scored one. In 2026-and-earlier exams, readers judged parts as Essentially correct, Partially correct or Incorrect (E, P, I) and derived a 0 to 4 question score. The May 2027 FRQs are worth 10 points each; consult the latest official sample rubric and scoring guidance for that format. No single phrase earns credit regardless of context.
Official sources
Course guidance reviewed in October 2026. Confirm the rules for your own exam session with the school and awarding body.
Questions parents ask
Is there a phrase that guarantees credit?
No. Credit depends on whether each part of the response is correct and complete for that question, in context.
Which exam year applies to my child?
Use the exam year they will actually sit. May 2027 is a revised, digital exam (42 MCQ and 4 FRQs); earlier years differ, so label every practice paper by year.