Free study guide — no login required
AP Statistics Study Guide (2026)
Last reviewed: 2026-06-10
AP Statistics is the College Board's introduction to data analysis, study design, probability, and statistical inference — the same content covered in a one-semester, non-calculus-based college statistics course. Unlike most math courses, success depends less on computation and more on interpretation: you'll spend the year describing distributions, critiquing how data was collected, and writing conclusions in context. The exam rewards students who can explain what a confidence interval means, not just compute one on a calculator.
The course is organized into nine units that build in a deliberate arc. Units 1-3 cover exploratory data analysis and data collection — histograms, regression, sampling methods, and experimental design. Units 4-5 develop the probability machinery, from random variables through the central limit theorem. Units 6-9 apply all of it to formal inference: confidence intervals and significance tests for proportions, means, chi-square procedures, and regression slopes.
This guide walks through every unit with its official CED exam weighting, the key topics the exam actually tests, and a study plan built on retrieval practice and spaced repetition. Whether you're starting in September or cramming in April, the structure below maps directly to what graders look for on exam day.
AP Statistics Exam Format
The AP Statistics exam is 3 hrs 15 min long and has 3 sections:
| Section | Format |
|---|---|
| Section I-A | 30 MCQs, no calculator (60 min) |
| Section I-B | 15 MCQs, calculator (45 min) |
| Section II | 6 FRQs (90 min) |
The exam runs three hours and is scored 1-5 from a composite of two equally weighted sections. Section I gives you 90 minutes for 40 multiple-choice questions with no guessing penalty, so answer everything. Section II gives you 90 minutes for six free-response questions: five standard FRQs plus an Investigative Task that stretches familiar methods into an unfamiliar setting and counts for a larger share of the section. A graphing calculator with statistical capabilities is allowed throughout, and a formula sheet with probability tables is provided.
FRQ scoring is holistic — each part is graded essentially complete, substantially complete, developing, or minimal — and graders look for four things on inference problems: name the procedure, check the conditions, show the mechanics, and state a conclusion in context linked to the p-value or interval. Never give a bare answer; always interpret in the context of the problem. On the Investigative Task, partial credit is generous, so attempt every part even when the scenario looks unfamiliar.
Who Should Take AP Statistics?
AP Statistics is one of the most broadly useful AP courses: statistics requirements appear in college majors from psychology and biology to business, nursing, political science, and data science, so a qualifying score frequently satisfies a required course rather than a generic elective. It's also accessible — the only real prerequisite is comfort with Algebra II, since the course uses almost no calculus. That makes it a strong choice for juniors and seniors who want a college-level math credit without taking AP Calculus, or alongside it. The challenge is different: precise vocabulary, careful condition-checking, and clear written justification matter more than algebraic skill.
AP Statistics Units: What to Study
Unit 1: Exploring One-Variable Data
15-23% of examThe heaviest-weighted unit covers describing a single variable. You'll distinguish categorical from quantitative variables, then represent quantitative data with dotplots, stemplots, histograms, and boxplots. Describing a distribution means addressing shape, center, variability, and unusual features — and choosing resistant measures (median, IQR) over the mean and standard deviation when distributions are skewed or contain outliers identified by the 1.5×IQR rule. The unit ends with measures of position: percentiles, z-scores, and the effects of linear transformations. Density curves lead into the normal distribution, where you'll use the empirical rule and calculator functions like normalcdf and invNorm to find areas and boundary values. Exam questions love comparison prompts — practice comparing two boxplots using full sentences with context.
Key topics
- Histograms, boxplots, and stemplots
- Shape, center, variability, outliers
- Mean vs. median resistance
- Standard deviation and IQR
- 1.5×IQR outlier rule
- Percentiles and z-scores
- Normal distribution and empirical rule
Unit 2: Exploring Two-Variable Data
5-7% of examThis short but conceptually dense unit handles relationships between two variables. For two categorical variables, you'll build two-way tables and compare marginal and conditional relative frequencies to detect association. For two quantitative variables, scatterplots get described by direction, form, strength, and unusual points, quantified by the correlation r. The core skill is least-squares regression: interpreting slope and y-intercept in context, computing and interpreting residuals, reading residual plots to judge whether a linear model is appropriate, and explaining r-squared as the percent of variation in y accounted for by the model. Expect questions on influential points, the danger of extrapolation, and the mantra that correlation does not imply causation. Reading regression output from computer printouts is a recurring exam task.
Key topics
- Scatterplots: direction, form, strength
- Correlation coefficient r
- Least-squares regression line
- Residuals and residual plots
- Interpreting slope in context
- r-squared (coefficient of determination)
- Outliers, leverage, and extrapolation
Unit 3: Collecting Data
12-15% of examUnit 3 asks how data should be gathered before it's analyzed. You'll contrast observational studies with experiments and learn why only well-designed experiments support cause-and-effect conclusions. Sampling methods include the simple random sample, stratified, cluster, and systematic sampling, along with the biases that wreck surveys: undercoverage, nonresponse, voluntary response, and question wording. Experimental design centers on the principles of comparison, random assignment, control, and replication, implemented through completely randomized, randomized block, and matched pairs designs. Confounding variables and the placebo effect explain why blinding matters. The exam's favorite question here is scope of inference: random sampling lets you generalize to a population, while random assignment lets you infer causation — and you must keep those two ideas separate.
Key topics
- Simple random, stratified, cluster sampling
- Sources of bias in surveys
- Observational study vs. experiment
- Confounding variables
- Completely randomized and block designs
- Matched pairs design
- Scope of inference
Unit 4: Probability, Random Variables, and Probability Distributions
10-20% of examThe probability unit supplies the engine for everything that follows. It opens with simulation and the law of large numbers, then builds the formal rules: complement, addition, and multiplication rules, conditional probability, and the definitions of independent versus mutually exclusive events — a distinction the exam tests relentlessly. Discrete random variables come next: building probability distributions, computing expected value and standard deviation, and finding the mean and variance of sums, differences, and linear transformations of random variables. The unit closes with two named distributions you must recognize from context: the binomial (fixed number of independent trials, BINS conditions) and the geometric (trials until first success). Calculator commands like binompdf and binomcdf handle the arithmetic, but setup and interpretation are on you.
Key topics
- Conditional probability
- Independent vs. mutually exclusive events
- Expected value of random variables
- Combining random variables
- Binomial distribution and BINS conditions
- Geometric distribution
- Simulation and law of large numbers
Unit 5: Sampling Distributions
7-12% of examThis is the conceptual hinge of the course: the idea that a statistic computed from a random sample is itself a random variable with its own distribution. You'll study the sampling distribution of a sample proportion — centered at p with standard deviation shrinking as n grows — and of a sample mean, where the central limit theorem guarantees approximate normality for large samples regardless of the population's shape. Key conditions debut here: the 10% condition for independence when sampling without replacement and the Large Counts condition (np and n(1−p) at least 10) for normality. The unit also covers biased versus unbiased estimators and sampling distributions for differences of two proportions or two means. Master this unit and Units 6-9 become applications of one repeated idea.
Key topics
- Sampling distribution of sample proportion
- Sampling distribution of sample mean
- Central limit theorem
- 10% condition
- Large Counts condition
- Unbiased estimators
- Differences of proportions and means
Unit 6: Inference for Categorical Data: Proportions
12-15% of examFormal inference begins with proportions. You'll construct and interpret one-sample z-intervals, understanding that the confidence level describes the long-run capture rate of the method, not the probability a particular interval contains p. Significance testing introduces the full framework: null and alternative hypotheses, the p-value as the probability of evidence at least this extreme assuming the null is true, and conclusions that compare p to a significance level alpha. The unit defines Type I and Type II errors, the power of a test, and the factors that increase power. Two-proportion z-intervals and z-tests extend everything to comparing groups. The graded template — state, plan (check random, 10%, Large Counts), do, conclude in context — is exactly what FRQ rubrics award points for.
Key topics
- One-proportion z-interval and z-test
- Interpreting confidence level vs. interval
- Null and alternative hypotheses
- P-value interpretation
- Type I and Type II errors
- Power of a test
- Two-proportion z-procedures
Unit 7: Inference for Quantitative Data: Means
10-18% of examInference for means swaps the z-distribution for the t-distribution, needed because the population standard deviation is estimated by the sample's. You'll work with degrees of freedom, one-sample t-intervals and t-tests, and the conditions: random sample, 10% condition, and Normal/Large Sample — satisfied by a roughly normal population, n of at least 30, or a sample graph showing no strong skew or outliers. The unit's biggest trap is choosing between paired and two-sample procedures: paired t-tests analyze the differences within matched pairs or repeated measurements on the same subjects, while two-sample t-tests compare means of independent groups. Exam FRQs frequently describe a study design and make you identify which procedure applies before any calculation, so practice that diagnosis step deliberately.
Key topics
- t-distribution and degrees of freedom
- One-sample t-interval and t-test
- Normal/Large Sample condition
- Paired t-procedures
- Two-sample t-test for means
- Choosing paired vs. two-sample
- Standard error of the mean
Unit 8: Inference for Categorical Data: Chi-Square
2-5% of examChi-square procedures handle categorical variables with more than two categories. The chi-square goodness-of-fit test checks whether one categorical variable's distribution matches a claimed model — like whether a die is fair or whether candy colors match the company's stated percentages. For two-way tables, you'll choose between the test for homogeneity (samples from multiple populations, comparing distributions) and the test for independence (one sample, two variables, testing association) — a distinction that hinges entirely on how the data was collected. Mechanics include computing expected counts, the chi-square statistic as a sum of (observed minus expected) squared over expected, degrees of freedom, and the Large Counts condition that all expected counts be at least 5. A follow-up skill: identifying which cells contribute most to the statistic.
Key topics
- Chi-square goodness-of-fit test
- Test for homogeneity
- Test for independence
- Expected counts calculation
- Degrees of freedom for two-way tables
- Large Counts condition for chi-square
- Component contributions to chi-square
Unit 9: Inference for Quantitative Data: Slopes
2-5% of examThe final unit closes the loop with Unit 2 by treating the least-squares slope b as a statistic with its own sampling distribution. You'll construct t-intervals for the true slope beta and run t-tests of the null hypothesis that beta equals zero, meaning no linear relationship between x and y in the population. Conditions follow the LINE acronym: Linear relationship, Independent observations, Normal distribution of residuals, and Equal standard deviation of y across all x values, plus random selection. The defining exam skill is reading computer regression output — identifying the slope estimate, its standard error, the t-statistic, and the p-value from a printout, then halving a two-sided p-value when the alternative is one-sided. Though lightly weighted, slope inference appears regularly on the Investigative Task.
Key topics
- Sampling distribution of the slope
- t-interval for slope
- t-test for slope (beta = 0)
- LINE conditions
- Reading computer regression output
- Standard error of the slope
- Connecting slope inference to scatterplots
How to Study for AP Statistics
Study the units in order, because the course is cumulative by design: you cannot interpret a p-value in Unit 6 without the sampling distributions of Unit 5, which depend on the probability rules of Unit 4. Budget the most time for Units 1, 3, 4, and 6-7 — together they carry the majority of the exam weight. As you finish each unit, build a one-page summary that includes every named condition (10%, Large Counts, Normal/Large Sample, LINE) and the exact sentence frames for interpretations: slope, r-squared, confidence interval, confidence level, and p-value. Graders award points for precise wording, so memorize the templates early.
Passive rereading does almost nothing for statistics; the skills are diagnostic. Use retrieval practice — close the notes and force yourself to identify the correct inference procedure from a bare scenario description, since procedure selection is the single highest-leverage exam skill. Schedule reviews with SM-2 spaced repetition: when you recall a concept correctly, the interval until the next review grows; when you miss it, the card resets to short intervals. MaxYourScore builds SM-2 scheduling into its unit quizzes, but flashcards work too — what matters is testing yourself at expanding intervals instead of rewatching videos the night before.
Six to eight weeks out, shift from learning to rehearsal. Work released FRQs under a 13-minute-per-question clock and score yourself against the actual rubrics, which the College Board publishes — you'll quickly see that 'conclude in context' costs more students points than any calculation. Take at least two full timed practice exams to build stamina for the 90-minute multiple-choice section. In the final week, drill your weakest unit's conditions and interpretations, re-derive the choice tree for all eight inference procedures from memory, and confirm your calculator's inference menu is second nature.
AP Statistics FAQ
Is AP Statistics hard?
AP Statistics is moderately difficult, but in an unusual way: the math rarely goes beyond Algebra II, yet the exam demands precise vocabulary, condition-checking, and written interpretation in context. Students who expect a computation course often struggle with the writing, while strong writers who are weaker at algebra frequently thrive. The hardest stretch for most students is Units 4-5, where probability and sampling distributions get abstract before inference makes them concrete again.
Is AP Statistics harder than AP Calculus?
Most students find AP Statistics less mathematically demanding than AP Calculus AB or BC — there are no limits, derivatives, or integrals, and a formula sheet is provided. But it isn't strictly easier: Statistics requires careful reading, justification in full sentences, and conceptual reasoning about randomness that many calculus-strong students find unfamiliar. If you excel at symbolic manipulation, calculus may feel more natural; if you prefer interpretation and applied reasoning, statistics usually will.
What percent do you need to get a 5 on AP Statistics?
The College Board doesn't publish fixed cutoffs, and the raw-score boundaries shift slightly each year as exams are equated for difficulty. Historically, earning roughly three-quarters of the available composite points has landed in 5 territory, with a 3 reachable at well under half. Because multiple-choice and free response each count for 50%, you can't ignore FRQ writing practice and expect the multiple-choice section to carry you.
Can you use a calculator on the AP Statistics exam?
Yes — a graphing calculator with statistical capabilities is allowed, and expected, on both the multiple-choice and free-response sections. Models like the TI-84 family handle the standard tests, intervals, and distribution calculations. You also receive a formula sheet and probability tables. One caution for free response: calculator commands alone don't earn full credit. Name the procedure and identify the values you used, rather than writing bare calculator syntax like normalcdf with unlabeled numbers.
What math do you need for AP Statistics?
Algebra II is the standard prerequisite, and it's genuinely sufficient — AP Statistics uses no calculus. You'll need comfort with linear equations (for regression), square roots and inequalities (for standard deviations and intervals), and careful arithmetic with formulas. The bigger demands are reading dense word problems, keeping notation straight (p versus p-hat, mu versus x-bar), and writing clear justifications. Many schools offer it to juniors and seniors concurrently with Precalculus or Calculus.
Ready to master AP Statistics?
Get all 9 unit videos, note packets, 180 quiz questions, 5 full-length practice exams, and a 24/7 Socratic AI tutor — $7.99/month with a 3-day free trial.