AP Statistics Free Response QuestionsFRQ Archive & Practice (2019 to 2025)

What do AP Statistics FRQs test?

Applying statistical reasoning to real contexts, not mechanical procedure recall.

The free response section accounts for 50% of the AP Statistics score, and it is the part that separates students who understand statistics from students who only memorize procedures. Where the 40 multiple choice questions sample breadth and conceptual recognition, the FRQs test whether you can set up the right procedure for a given context, check that its conditions are met, carry out the calculation correctly, and then interpret the result in plain English with reference to the specific study or data set in the prompt. No FRQ rewards procedure execution alone. The credited answer always includes an in context interpretation: what the interval means for this population, what the p value means for this claim, whether the study design supports a causal conclusion. The College Board's four statistical practices, particularly Statistical Argumentation (Skill 4), are tested on virtually every question and account for most points lost through incomplete or context free conclusions.

Standard FRQs versus the Investigative Task on AP Statistics

Five standard questions target focused procedures; one Investigative Task requires integrating concepts across the entire course.

Section II is always 6 questions: Q1 through Q5 are standard FRQs, each typically worth 4 points and targeting a specific statistical procedure or concept, while Q6 is the Investigative Task, the most complex and most heavily weighted question on the exam. The Investigative Task is unique to AP Statistics among all AP exams.

How are AP Statistics FRQs scored?

Analytic point rubrics with a holistic scoring scale: E (Essentially Correct), P (Partially Correct), and I (Incorrect) per part, then converted to a 4 point scale per question.

AP Statistics uses a distinctive rubric system. For each question, trained Readers score individual parts on a three level scale: E (Essentially Correct, full credit), P (Partially Correct, half credit), and I (Incorrect, no credit). The total E and P counts across a question's parts are then converted to a question score on a 0 to 4 point scale using a conversion table in the scoring guideline. This system means that partial credit accumulates across parts rather than being binary per question. A response that is Essentially Correct on three of five parts and Partially Correct on the other two earns more than a response that fully answers two parts and omits the rest. There is no penalty for attempting a part incorrectly. The Investigative Task uses the same E, P, I rubric across its sub parts but the conversion table reflects its larger total point value. The full year by year scoring guidelines, and how the FRQ raw total is weighted into the composite, are on the AP Statistics scoring guidelines page.

Worked example: how a real AP Statistics FRQ was scored

2023 Question 4, confidence interval for a proportion. Max 4 points, national mean approximately 1.8 points.

Question 4 from the 2023 AP Statistics exam asked students to construct and interpret a 95% confidence interval for a population proportion, one of the most common FRQ formats. It illustrates exactly which steps the rubric requires and which phrasings lose points. The rubric uses E, P, I scoring on each part, then converts the totals to a 0 to 4 scale.

1. Part A: State and check all required conditions for the procedure.

   Rubric: Essentially Correct requires stating the random condition, the independence condition (10% rule), and the large counts condition, each with explicit verification using the sample data. All three must appear with reference to the specific numbers in the problem.

   Earns the point: Random: the adults were randomly selected from the population. Independence: the sample of 1,000 is less than 10% of all adults. Large counts: the estimated number of successes is (1,000)(0.42) = 420 and estimated failures is (1,000)(0.58) = 580, both at least 10, so the large counts condition is met.

   Loses the point: Writing only that the sample is large enough without computing the estimated counts (420 and 580) earns Partially Correct at best. Omitting the 10% condition check earns I on that condition regardless of the other two.

2. Part B: Calculate the 95% confidence interval.

   Rubric: Essentially Correct requires the correct formula set up with correct values, the correct critical value (z* = 1.96 for 95%), the correct standard error, and correct interval arithmetic, either shown or produced by a correctly identified calculator command.

   Earns the point: 0.42 plus or minus 1.96 times the square root of (0.42 times 0.58 divided by 1,000), giving 0.42 plus or minus 0.031, so the interval is (0.389, 0.451).

   Loses the point: Using z* = 2 or z* = 1.645 instead of 1.96 earns Partially Correct. Reporting only the calculator output without identifying the procedure and inputs earns I on the setup portion. Dividing by n instead of the square root of n earns I on the standard error.

3. Part C: Interpret the interval in the context of this problem.

   Rubric: Essentially Correct requires explicit reference to the population parameter being estimated, the specific context of the study, and the use of confident language. The response must not use probability language about a fixed interval (saying there is a 95% probability the true proportion is in this interval is incorrect after the interval is computed).

   Earns the point: We are 95% confident that the true proportion of all adults in the population who have the characteristic is between 0.389 and 0.451.

   Loses the point: There is a 95% probability that the true proportion of adults who have the characteristic is in this interval. (Probability language about a fixed computed interval is incorrect and earns I on the interpretation.) Also losing: stating the interval without saying what parameter it estimates, or giving a correct interpretation that omits the specific context from the problem.

The AP Statistics FRQ rubric consistently rewards three things in sequence: stated and checked conditions with numbers from the problem, a correctly set up and computed procedure, and an interpreted conclusion using confident language and the specific context of the question. Missing any one of these steps costs a point. The most frequent loss across all three parts is the interpretation: students who compute the correct interval still earn Partially Correct if they drop the context or use probability language instead of confidence language. Students who answer all three parts in order and use the correct vocabulary, confident, parameter, and population, reliably earn full credit.

Common AP Statistics FRQ mistakes

1. 01

   Incomplete interpretation of confidence intervals

   Students construct the correct numerical interval but fail to interpret it in the context of the problem. A complete interpretation must name the population parameter being estimated, reference the specific study context, and use confidence language (we are 95% confident that...). Chief Reader Reports flag responses that either omit the context entirely, interpret the interval as a statement about the sample rather than the population, or use incorrect probability language (there is a 95% probability the true proportion falls in this interval), which is statistically wrong for a fixed computed interval.

   AP Statistics Chief Reader Reports 2022 to 2024, inference for proportions questions

2. 02

   Failing to state or check all three required conditions

   On every inference FRQ, students must verify three conditions: the random condition (data come from a random sample or random assignment), the independence condition (the 10% rule: the sample size is less than 10% of the population), and either the large counts condition (for proportions) or the normality condition (for means). Chief Reader Reports note that students routinely state only one or two conditions, or list the condition names without checking them against the actual sample values. Omitting or inadequately checking conditions costs points on every inference question because each condition is evaluated separately in the E, P, I rubric.

   AP Statistics Chief Reader Reports 2021 to 2025, inference procedure questions across Units 6, 7, and 8

3. 03

   Ambiguous or missing null and alternative hypotheses

   Students write the null and alternative hypotheses using Greek letter notation without defining what the parameter represents in the context of the problem, or they reverse the direction of the alternative hypothesis. Per the AP Statistics scoring guidelines, hypotheses must define the parameter in words tied to the specific research question. Writing H0: p = 0.5 without stating that p represents the true proportion of all adults in the survey who hold a particular view earns Partially Correct or Incorrect depending on the rubric year.

   AP Statistics Chief Reader Reports 2021 to 2024, significance test questions

4. 04

   Out of context conclusions after hypothesis tests

   Students state reject H0 or fail to reject H0 as the complete conclusion, without completing the sentence with a statement about the research question in the specific context of the study. The AP Statistics rubric requires a conclusion that restates the practical meaning: for example, there is sufficient evidence to conclude that the true proportion of voters who support the policy is greater than 0.5. A standalone reject H0 earns I on the conclusion part regardless of the correctness of the test procedure.

   AP Statistics Chief Reader Reports 2022 to 2025, hypothesis test conclusion parts

5. 05

   Confusing paired and two sample procedures

   When a study design involves matched pairs (each subject measured twice, or each pair of subjects matched on a characteristic), the correct procedure is a one sample t test or t interval on the differences, not a two sample t procedure on the original values. Chief Reader Reports consistently cite students who run a two sample t test on paired data, which is both the wrong procedure and a condition violation, since the two measurements are not independent. Recognizing pairing in the problem context before selecting a procedure is a prerequisite for earning credit.

   AP Statistics Chief Reader Reports 2021, 2023, 2024, Unit 7 paired versus two sample questions

6. 06

   Treating the Investigative Task as five disconnected standard FRQs

   The Investigative Task is built around a single extended scenario, and the sub parts are designed so that later parts build on findings from earlier ones. Chief Reader Reports identify a pattern in which students answer each sub part in isolation without referencing the results of prior parts, missing the cumulative statistical argument the task requires. Strong Investigative Task responses explicitly connect conclusions from earlier sub parts to the setup and interpretation of later ones, demonstrating the integrated reasoning across multiple statistical skills that distinguishes a score of Essentially Correct from Partially Correct on the task as a whole.

   AP Statistics Chief Reader Reports 2022 to 2025, Investigative Task Q6 feedback

How to practice AP Statistics FRQs effectively

Timed reps under exam conditions, then self score against the official rubric, part by part.

The highest return AP Statistics FRQ practice is not reading through released questions but working one under timed conditions and then grading yourself line by line against that year's official scoring guideline from College Board. The archive above links every year's official booklet through the College Board past exam questions page, where the matching scoring guidelines are also posted. Work one standard FRQ in about 12 minutes, score it against the E, P, I rubric in the guideline, and note every part where you earned P instead of E or I instead of P. After three or four sessions the pattern in your losses becomes specific and almost always traces to conditions not checked with numbers, hypothesis parameters not defined in context, or conclusions dropped before the contextual statement. For the Investigative Task, work the full Q6 in about 25 minutes without interruption to simulate the sustained reasoning it requires. Comparing your wording to the sample responses in the scoring guidelines reveals the exact phrasing that earns Essentially Correct versus Partially Correct.

1. 1

   Use the 10 minute reading period to read all 6 questions and plan your approach. Identify which standard FRQs you can answer most confidently and start with those to bank secure points before tackling the Investigative Task.

2. 2

   For every inference question, write your conditions before touching the calculation. State the random condition, the 10% condition, and the large counts or normality condition, each checked with numbers from the problem. A missing condition costs a point regardless of how accurate the interval or test is.

3. 3

   State hypotheses in words tied to the specific context. Writing H0: p = 0.5 earns no credit unless you define p as the true proportion of all adults in the survey who have a specific characteristic. Define the parameter every time.

4. 4

   End every significance test with a full conclusion in context. After stating whether you reject or fail to reject H0, complete the sentence: there is (or is not) sufficient evidence to conclude that the true proportion of X is greater than (or different from) Y in this population.

5. 5

   For the Investigative Task, treat it as one extended argument, not five separate problems. Before starting sub part C or D, re read your conclusions from sub parts A and B. Explicitly reference those findings in later parts to show the cumulative reasoning the rubric rewards.

6. 6

   On paired data problems, recognize pairing before you choose a procedure. If each observation in one group is matched to a specific observation in the other group, compute the differences and use a one sample t procedure on those differences rather than a two sample t procedure.

7. 7

   Practice writing interpretations aloud or in writing for every confidence interval and every test conclusion. The required phrasing, we are 95% confident that the true proportion of all X who Y is between A and B, is a specific learned formula that earns points consistently when used correctly and loses them reliably when paraphrased incorrectly.

AP Statistics FRQ FAQ

More AP Statistics resources