AP · Analysis & Prediction

2025 AP AP Statistics Past Paper: Analysis & Answers

The 2025 AP Statistics Free-Response section provided a balanced assessment spanning data exploration, experimental design, binomial and discrete probability distributions, hypothesis testing for proportions and means, and an investigative task evaluating effect sizes (Cohen's d). Candidate performance revealed significant recurring issues in writing rigorous statistical justifications, properly referencing population parameters, and avoiding deterministic language in inference conclusions.

Updated: 18 Aug 2026

Analysis Sample paper

Difficulty 3.4/5

Total marks

Duration

90 mins

Most tested

Inference for Quantitative Data: Means (Unit 7)

Paper breakdown

Section II: Free-Response: 24 marks · 90 mins

Top chapters

Unit 7: Inference for Quantitative Data: Means · 6.5 marksUnit 4: Probability, Random Variables, and Probability Distributions · 5.5 marksUnit 1: Exploring One-Variable Data · 4 marksUnit 3: Collecting Data · 4 marksUnit 6: Inference for Categorical Data: Proportions · 4 marks

Overall Verdict & Exam Profile

The 2025 AP Statistics Free-Response examination proved to be moderately challenging with an average student score of \(10.26\) out of \(24\) points across the six questions. Question 2 (sampling design) had the highest average performance at \(2.10/4\), whereas Question 5 (multi-focus discrete probability and \(t\)-test concepts) and Question 3 (binomial probability modeling) presented the highest hurdles, averaging \(1.43\) and \(1.55\) points respectively. Question 6 (Investigative Task) introduced Cohen's \(d\) effect size formula to test students' ability to conceptualize practical versus statistical significance.

Where the Marks Were Won and Lost

Marks were readily achieved on routine calculations—such as computing the range of a combined distribution, the mean of a discrete random variable, and carrying out arithmetic test statistics. However, substantial marks were lost across multiple questions due to incomplete justifications and missing context:

Unit 1 (Question 1): Many students failed to explicitly use comparative vocabulary (e.g., higher, lower) when analyzing boxplot centers and spreads, or attempted to deduce shape/modality from boxplots where it cannot be verified.
Unit 3 (Question 2): When describing stratified sampling implementation, candidates frequently omitted crucial practical details, such as properly mixing slips in a hat or defining the precise mapping of random numbers to field regions.
Unit 4 (Question 3 & 5): Defining random variables was a common pitfall; students frequently described \(X\) as 'rock songs' rather than specifying the count of rock songs in a defined sample of \(20\) trials.
Units 6 & 7 (Questions 4, 5, 6): In hypothesis testing, students consistently lost points by using deterministic language (e.g., 'proves the null/alternative') rather than non-definitive phrasing ('convincing statistical evidence'), and by expressing hypotheses using sample statistics rather than population parameters.

Key Pitfalls & Strategy for High Scores

To maximize scores on future AP Statistics exams:

Context is Non-Negotiable: Always include the variable, units, and explicit reference to the population (e.g., 'the true mean number of bedrooms in newly built houses in 2024') rather than generic labels.
Master Random Variable Definitions: Explicitly state the distribution name, trials \(n\), and probability \(p\) whenever setting up binomial probability models.
Differentiate Statistical vs. Practical Significance: The Investigative Task emphasized that a tiny \(p\)-value guarantees statistical significance but does not automatically mean the observed effect is large in real-world magnitude.

Ready to test yourself?

Turn these notes into exam-style practice. Get unlimited AI questions on this topic with instant marking and explanations.

Practise This Topic

Charts

Marks by chapter

See where the marks were concentrated so revision time goes to the highest-value topics.

Question type mix

Compare the mark share of each paper section and question type.

Short Free Response (Q1-Q5)20 marks · 5 questions
Investigative Task (Q6)4 marks · 1 questions

Mark accessibility

Estimate which marks were basic, mid-level, or high-difficulty.

75%within easy or medium reach

easy: 7 marksmedium: 11 markshard: 6 marks

Difficulty trend

Compare difficulty across recent years.

Command word frequency

Spot common command words so answers match the expected response style.

Skill weighting

Shows the skill mix this paper tested most heavily.

Study ROI

Bigger bubbles recur more often; higher bubbles carry more marks, helping you rank revision priorities.

Inference for Proportions and Means (Significance Tests & CIs)

10.5 marks · Difficulty 3.2 · recurrence 100%

Probability, Binomial Distributions & Expected Values

5.5 marks · Difficulty 3.5 · recurrence 95%

Experimental Design & Sampling Procedures

4 marks · Difficulty 2.5 · recurrence 90%

Exploring One-Variable Data & Boxplot Comparisons

4 marks · Difficulty 2.8 · recurrence 85%

Time vs marks

Compare marks with suggested time allocation to plan exam pacing.

marksmins

Next-year prediction

Topics worth watching next year, with the reason shown directly below each bar.

Unit 8: Chi-Square Tests (Independence / Homogeneity / Goodness of Fit)85%

Unit 8 received zero FRQ representation in 2025 after appearing in 2024, making it highly probable for the next free-response rotation.

Unit 2: Two-Variable Quantitative Data & Linear Regression Residuals80%

Bivariate regression and residual plot analysis were absent from the 2025 FRQs and are standard alternating staples.

Unit 9: Inference for Linear Regression Slopes65%

Unit 9 has not appeared as a dedicated FRQ since 2023, making it a prime candidate for Question 4, 5, or 6.

Cumulative marks ladder

The line is your running mark total question by question; dashed lines are the estimated grade cut-offs. See which question the line crosses your target grade at, so you know how far you must answer cleanly and which questions decide a band.

Topic-year heatmap

Compare topic weight by year to spot recurring and returning areas.

May 2023

May 2024

May 2025

Unit 1: Exploring One-Variable Data

Unit 2: Exploring Two-Variable Data

2.5

Unit 3: Collecting Data

Unit 4: Probability, Random Variables, and Probability Distributions

5.5

Unit 5: Sampling Distributions

Unit 6: Inference for Categorical Data: Proportions

Unit 7: Inference for Quantitative Data: Means

6.5

Unit 8: Inference for Categorical Data: Chi-Square

Unit 9: Inference for Quantitative Data: Slopes

1.5

Examiner insights

Official report

Common pitfalls

Using definitive or proof language (e.g., 'we prove that', 'accept the null') instead of non-definitive phrasing ('convincing statistical evidence to suggest') when concluding significance tests.
Defining hypotheses in terms of sample statistics (e.g., p-hat or x-bar) rather than true population parameters (p or mu).
Omitting clear physical implementation details when describing random sampling procedures (e.g., forgetting to specify that slips of paper in a hat must be mixed thoroughly).
Failing to explicitly use comparative terms (e.g., greater than, lower spread) when comparing two distributions from graphical displays.
Stating that the distribution of a random variable is 'distributed randomly' or 'normal' instead of specifying the binomial distribution with parameters n and p.

Misconceptions

Believing that distribution shape or modality (e.g., unimodal, bell-shaped) can be definitively deduced from a standard five-number boxplot.
Confusing Type I and Type II errors, specifically forgetting that a Type I error is conditional upon the null hypothesis being true in reality.
Calculating the median of a combined dataset by simply averaging the medians of the two separate datasets.
Confusing statistical significance (measured by p-value) with practical real-world importance (measured by effect size / Cohen's d).

Where marks are lost

Failing to check both parts of the independence condition (SRS and the 10% condition with correct population context) in Question 4 Section 2.
Using calculator function syntax (e.g., binomcdf(20, 0.1, 3)) without clearly defining all parameter inputs and boundary values.
Omitting the absolute value or forgetting to square individual sample standard deviations when calculating pooled standard deviation for Cohen's d in Question 6.
Stating conclusions without explicitly connecting the numerical p-value comparison to the given significance level alpha.

Know what the examiners want. Now go get those marks

thinka is an AI practice app used by 100,000+ students to drill questions and get instant feedback on every answer. You already know the high-frequency topics and where marks are lost. Now practise those specifically on thinka and turn this analysis into actual marks.

Start Practising Free See pricing

Know what's tested? Drill those topics on thinka. AI questions, instant marking.

Start Practising Free