Review AP Statistics Unit 5 to understand how and why sample statistics vary from sample to sample, and how the normal distribution and Central Limit Theorem let you model that variation precisely. This unit is the bridge between probability and every inference procedure in Units 6 through 9.
Use the topic guides, key terms, and practice questions available for this unit to build fluency with sampling distribution formulas and conditions before moving to inference.
When you take a sample from a population, the statistic you calculate will not match the true parameter exactly, and it will not match the statistic from a different sample either. Unit 5 explains why that happens and gives you the mathematical tools to describe and quantify that variation.
Two samples from the same population will almost always produce different statistics. That variation can be random (chance) or non-random (bias from a flawed design). Recognizing the source of variation is the first step before any inference.
A normal distribution is symmetric and bell-shaped, described by mean mu and standard deviation sigma. Area under the curve over an interval equals the probability a value falls there. Z-scores and calculator functions like normalcdf and invNorm let you find those areas precisely.
For proportions, the Large Counts condition (np >= 10 and n(1-p) >= 10) must hold. For means, either the population is normal or n >= 30 by the CLT. The 10% condition (n < 10% of population) justifies treating observations as independent when sampling without replacement.
Every confidence interval and hypothesis test in Units 6 through 9 rests on knowing the center, spread, and shape of a sampling distribution. If you can state the mean and standard deviation of the distribution of p-hat or x-bar, and justify a normal model, you can calculate any probability about a sample result and set up any inference procedure.
Introduces sampling variability: statistics from repeated samples of the same population differ due to random chance or non-random bias. Establishes why conclusions from a single sample carry uncertainty.
Reviews the normal distribution as a model for continuous random variables. Covers z-scores, area under the curve as probability, and using normalcdf and invNorm to find probabilities and boundary values.
Defines sampling distributions and states the CLT: with independent observations and sufficiently large n, the sampling distribution of x-bar is approximately normal. Introduces randomization distributions via simulation.
Distinguishes biased from unbiased estimators. The sample mean and sample proportion are unbiased. Larger samples reduce variability but cannot correct bias from a flawed sampling method.
Describes the center (p), spread (sqrt(p(1-p)/n)), and shape (approximately normal when Large Counts holds) of the sampling distribution of p-hat. Applies the 10% condition for without-replacement sampling.
Extends proportion sampling distributions to two independent groups. Mean is p1 minus p2, standard deviation uses both group formulas, and normality requires all four Large Counts checks to pass.
Describes the center (mu), spread (sigma/sqrt(n)), and shape of the sampling distribution of x-bar. Normal when population is normal; approximately normal for any population when n >= 30 by the CLT.
Extends mean sampling distributions to two independent groups. Mean is mu1 minus mu2, standard deviation is sqrt(sigma1^2/n1 + sigma2^2/n2), and normality requires both populations normal or both n >= 30.
Open this guide for a closer review of the topic.
Open this guide for a closer review of the topic.
Open this guide for a closer review of the topic.
Open this guide for a closer review of the topic.
Open this guide for a closer review of the topic.
Open this guide for a closer review of the topic.
This snapshot uses Fiveable practice activity to show where students tend to miss questions and which review moves are worth prioritizing first.
Across 7.7k multiple-choice practice attempts for this unit.
Practice activity included in this snapshot.
Across 45 scored free-response attempts for this unit.
Review Biased and Unbiased Point Estimates with attention to how the concept appears in AP-style source and evidence questions.
Review Sampling Distributions for Sample Proportions with attention to how the concept appears in AP-style source and evidence questions.
Review Sampling Distributions for Differences in Sample Means with attention to how the concept appears in AP-style source and evidence questions.
Review Sampling Distributions for Sample Means with attention to how the concept appears in AP-style source and evidence questions.
Every sample statistic is subject to sampling variability: repeated samples from the same population produce different values of x-bar or p-hat. That variation may be random (chance) or non-random (systematic bias from poor design). A single sample always carries uncertainty about the true population parameter.
A normal distribution is fully described by its mean mu and standard deviation sigma. The area under the curve over any interval equals the probability a value falls there. Use z = (x - mu)/sigma to standardize, then a z-table or normalcdf to find area. Use invNorm to work backward from a given probability to a boundary value.
| Task | Formula or tool | Example |
|---|---|---|
| Find P(X < a) | normalcdf(-inf, a, mu, sigma) | P(X < 78) with mu=70, sigma=8 |
| Find P(a < X < b) | normalcdf(a, b, mu, sigma) | P(65 < X < 85) |
| Find P(X > b) | normalcdf(b, inf, mu, sigma) | P(X > 90) |
| Find boundary for given area | invNorm(area, mu, sigma) | Find x where P(X < x) = 0.25 |
| Standardize a value | z = (x - mu)/sigma | z = (82 - 70)/8 = 1.5 |
A sampling distribution of a statistic is the distribution of that statistic across all possible samples of size n from a given population. The Central Limit Theorem (CLT) states that when n is sufficiently large and observations are independent, the sampling distribution of the sample mean is approximately normal regardless of the population's shape. A randomization distribution is a simulated version of this, generated by repeatedly reassigning response values to treatment groups.
A point estimator is a sample statistic used to estimate a population parameter. An estimator is unbiased if its sampling distribution is centered at the true parameter value, meaning the average of all possible estimates equals the parameter. The sample mean x-bar and sample proportion p-hat are both unbiased estimators. Increasing sample size reduces variability in the estimator but does not fix bias if the sampling method is flawed.
When you take random samples of size n from a population with proportion p, the sample proportion p-hat has a predictable sampling distribution. Its mean equals p (unbiased), its standard deviation equals sqrt(p(1-p)/n), and it is approximately normal when the Large Counts condition holds. The 10% condition justifies using this formula when sampling without replacement.
When comparing two independent groups, the statistic of interest is p-hat1 minus p-hat2. Its sampling distribution has mean equal to p1 minus p2 and standard deviation equal to sqrt(p1(1-p1)/n1 + p2(1-p2)/n2). It is approximately normal when all four Large Counts checks pass: n1p1 >= 10, n1(1-p1) >= 10, n2p2 >= 10, and n2(1-p2) >= 10.
For a numerical variable sampled from a population with mean mu and standard deviation sigma, the sampling distribution of x-bar has mean mu_x-bar = mu and standard deviation sigma_x-bar = sigma/sqrt(n). It is exactly normal if the population is normal, and approximately normal for any population shape when n >= 30 by the CLT. Use z = (x-bar minus mu)/(sigma/sqrt(n)) to find probabilities.
When comparing two independent groups on a numerical variable, the statistic x-bar1 minus x-bar2 has a sampling distribution centered at mu1 minus mu2 with standard deviation sqrt(sigma1^2/n1 + sigma2^2/n2). The distribution is approximately normal when both populations are normal or when both sample sizes are at least 30.
| Statistic | Mean of sampling distribution | Standard deviation of sampling distribution | Normal model condition |
|---|---|---|---|
| p-hat | p | sqrt(p(1-p)/n) | np >= 10 and n(1-p) >= 10 |
| p-hat1 minus p-hat2 | p1 minus p2 | sqrt(p1(1-p1)/n1 + p2(1-p2)/n2) | All four large counts >= 10 |
| x-bar | mu | sigma/sqrt(n) | Population normal OR n >= 30 |
| x-bar1 minus x-bar2 | mu1 minus mu2 | sqrt(sigma1^2/n1 + sigma2^2/n2) | Both populations normal OR both n >= 30 |
Try AP-style multiple-choice questions and written prompts after you review the notes.
| Term | Definition |
|---|---|
| Parameter | A numerical value that describes a characteristic of a population, such as mu or p; the target that sample statistics are used to estimate. |
| Sample Statistic | A numerical value calculated from sample data, such as x-bar or p-hat, used as a point estimate of the corresponding population parameter. |
| Unbiased Estimator | A statistic whose sampling distribution is centered at the population parameter it estimates; x-bar is unbiased for mu and p-hat is unbiased for p. |
| Standard Error | The standard deviation of a sampling distribution; for x-bar it equals sigma/sqrt(n), and for p-hat it equals sqrt(p(1-p)/n). Larger n produces a smaller standard error. |
| Z-scores | A standardized value calculated as (data value minus mean) divided by standard deviation; used to find probabilities under a normal curve. |
| Empirical Rule | In a normal distribution, approximately 68% of values fall within 1 standard deviation of the mean, 95% within 2, and 99.7% within 3. |
| Large Counts Condition | The requirement that np >= 10 and n(1-p) >= 10 for the sampling distribution of p-hat to be approximately normal; all four analogous counts must hold for differences in proportions. |
| ten percent condition | The requirement that sample size be less than 10% of the population size to treat observations as independent when sampling without replacement. |
| Sample Proportion | p-hat; the ratio of successes in a sample to total observations, used to estimate the population proportion p. |
| Sample Mean | x-bar; the average of values in a sample, used to estimate the population mean mu. Its sampling distribution has mean mu and standard deviation sigma/sqrt(n). |
| Normality | The property of a distribution being symmetric and bell-shaped; required (or approximately satisfied) before using normal probability calculations for sampling distributions. |
| randomization distribution | A simulated sampling distribution created by repeatedly and randomly reallocating response values to treatment groups in a randomized experiment; used to assess whether an observed result is unusual. |
| Bias | Systematic error in an estimator or sampling method that consistently pushes estimates away from the true parameter value; not reduced by increasing sample size. |
When finding probabilities about x-bar, the standard deviation of the sampling distribution is sigma/sqrt(n), not sigma. Using sigma gives the spread of individual values, not the spread of sample means. Always divide by sqrt(n) when working with x-bar.
The normal model for p-hat requires np >= 10 and n(1-p) >= 10. For x-bar, you need either a normal population or n >= 30. For differences, all conditions must hold for both groups. Stating a normal model without verifying conditions loses credit on the exam.
The 10% condition (n < 10% of population) addresses independence when sampling without replacement. The Large Counts condition (np >= 10 and n(1-p) >= 10) addresses whether the normal model applies. These are separate checks that serve different purposes.
A larger sample size reduces variability in an estimator but does not remove bias. If the sampling method systematically excludes part of the population, the estimator will still be off-center no matter how large n gets.
A probability statement about p-hat or x-bar must reference the specific population and variable described in the problem. Writing P(x-bar > 53) = 0.067 without saying what that means in context is incomplete on a free-response question.
Free-response questions on sampling distributions routinely require you to check and justify conditions before proceeding. For proportions, state and verify the Large Counts condition and 10% condition explicitly. For means, state whether the population is normal or invoke the CLT with n >= 30. Skipping this step or stating conditions without numbers from the problem typically results in lost credit.
A common task is computing the probability that a sample statistic falls in a given range. You must use the standard deviation of the sampling distribution (sigma/sqrt(n) for x-bar, sqrt(p(1-p)/n) for p-hat), not the population standard deviation. Show the z-score formula with numbers substituted, then state the probability with a calculator result and interpret it in context.
Exam questions expect you to connect numerical results to the real-world scenario described. A probability about x-bar should be stated as the likelihood that the sample mean of a specific variable in a specific population falls in a given range, not just as a decimal. Similarly, the mean and standard deviation of a sampling distribution should be described with units and context, not just as formulas.
Open the individual guides for Unit 5 when you want a closer review of one topic.
browse guidesPractice free-response reasoning and compare your answer with scoring guidance.
practice FRQsWatch past review streams filtered to Unit 5 when you want a video walkthrough.
open videosUse unit cheatsheets for a quick visual review after you work through the notes.
open cheatsheetsEstimate your broader AP score goal after you review the course and exam format.
open calculatorAP Stats Unit 5 covers 8 topics on sampling distributions: the normal distribution, the Central Limit Theorem, biased and unbiased point estimates, sampling distributions for sample proportions, differences in sample proportions, sample means, and differences in sample means. The unit builds the statistical foundation you need for inference. Here's the full topic list: - 5.1 Introducing Statistics: Why Is My Sample Not Like Yours? - 5.2 The Normal Distribution, Revisited - 5.3 The Central Limit Theorem - 5.4 Biased and Unbiased Point Estimates - 5.5 Sampling Distributions for Sample Proportions - 5.6 Sampling Distributions for Differences in Sample Proportions - 5.7 Sampling Distributions for Sample Means - 5.8 Sampling Distributions for Differences in Sample Means See all the matched practice at AP Stats Unit 5.
AP Stats Unit 5 makes up 7-12% of the AP exam. That weight covers sampling distributions, the normal distribution, and the Central Limit Theorem. These concepts are also the backbone of Units 6-9, so understanding them well pays off across a much larger portion of the exam than that percentage suggests.
The AP Stats Unit 5 progress check on AP Classroom includes both MCQ and FRQ parts drawn from all 8 topics in the unit. MCQ questions test the normal distribution, Central Limit Theorem, and sampling distributions for proportions and means. FRQ prompts typically ask you to identify, set up, and interpret a sampling distribution in context. The progress check pulls heavily from topics 5.2 through 5.8, so make sure you're comfortable calculating probabilities using the normal distribution and explaining why the Central Limit Theorem applies for a given sample size. For matched practice problems that mirror the progress check format, visit AP Stats Unit 5.
AP Stats Unit 5 FRQs most often come from the normal distribution, the Central Limit Theorem, and sampling distributions for sample proportions and means. A typical prompt gives you a real-world scenario and asks you to describe the shape, center, and spread of a sampling distribution, then calculate a probability or explain what the Central Limit Theorem guarantees. To practice effectively, work through each step out loud: state conditions, show the formula, calculate, and interpret in context. That last step, writing a sentence that ties your number back to the scenario, is where most points are lost. You'll find FRQ-style practice problems organized by topic at AP Stats Unit 5.
The best place to find AP Stats Unit 5 practice questions, including multiple-choice and practice test sets, is AP Stats Unit 5. That page organizes MCQ and FRQ practice by topic, covering the normal distribution, Central Limit Theorem, and all four sampling distribution types (proportions, differences in proportions, means, and differences in means). For a focused practice test experience, work through topic-by-topic MCQs first to spot gaps, then move to full FRQ prompts. Targeting topics 5.3, 5.5, and 5.7 first gives you the highest return since those show up most on both the progress check and the actual AP exam.
Start AP Stats Unit 5 by locking in the normal distribution (topic 5.2) before anything else, since every later topic builds on it. Then work through the Central Limit Theorem (5.3) carefully and practice explaining in plain English why a large enough sample size makes the sampling distribution approximately normal. Here's a practical study sequence: 1. Review the normal distribution and practice z-score probability calculations. 2. Study the Central Limit Theorem and know the conditions: random sample, independence, and large enough n. 3. Work topics 5.4-5.8 in order, sketching the sampling distribution (shape, mean, standard deviation) for each scenario before calculating. 4. For every practice problem, write a one-sentence interpretation of your answer in context. The most common mistake is skipping the conditions check. On the AP exam, stating and verifying conditions is worth points on its own. Visit AP Stats Unit 5 for topic-by-topic practice to reinforce each step.