Sample Size in AP Statistics

In AP Statistics, sample size (n) is the number of individuals or observations in a sample. A larger n shrinks the variability of a sampling distribution (σ/√n), narrows confidence intervals, and increases a test's power, but it does not fix bias.

What is Sample Size?

Sample size, written n, is how many observations you actually collected. If you survey 400 teens, n = 400. It is not the size of the population, and it is not how many samples you took. It's the count inside one sample.

Why does n matter so much? It sits in the denominator of almost every standard deviation formula for a sampling distribution. The standard deviation of x̄ is σ/√n, and for p̂ it's √(p(1−p)/n). Bigger n means less sample-to-sample variability, so your statistic lands closer to the true parameter more often. The square root is the catch. To cut variability in half, you need four times the sample size. Sample size also decides whether you can trust a normal model. That's where the Central Limit Theorem, the Large Counts condition, and the 10% condition come in.

Why Sample Size matters in AP® Statistics

Sample size shows up in nearly every inference topic. It starts with the idea of collecting data (1.1) and runs through sampling distributions and the Central Limit Theorem (2.12, 4.1, 3.9, 4.6). It then carries into confidence intervals for proportions and means (3.3, 3.4, 3.11, 4.2, 4.3, 4.8), test setup (4.4), and Type I/Type II errors and power (3.8). It is the one variable you, the researcher, actually control. So the exam keeps asking what happens when n changes. How does the interval width change? What about the power or the shape of the sampling distribution? If you understand n, you understand why inference works at all.

How Sample Size connects across the course

Sampling Distributions and the Central Limit Theorem (Units 2, 4)

The CLT says that as n grows, the sampling distribution of x̄ looks more and more normal, even when the population is skewed. Think of n as the dial that turns a lumpy population into a smooth bell curve of sample means. That's why a heavily right-skewed population, like household electricity use, needs a large sample (the common rule of thumb is n ≥ 30) before you can use a normal model.

Margin of Error and Confidence Intervals (Units 3, 4)

Margin of error is critical value times standard error, and standard error shrinks as √n grows. So a bigger sample gives a narrower interval at the same confidence level. Because of the square root, halving the margin of error takes four times as many observations. You can also run this backward and solve for the minimum n that hits a target margin of error.

Power of a Test and Type II Error (Unit 3)

Power is the chance a test correctly rejects a false null hypothesis. Larger samples make the sampling distribution tighter, so a real difference is easier to spot. More n means more power and a lower chance of a Type II error, with α staying the same.

Conditions for Inference: 10% and Large Counts (Units 2, 3, 4)

Sample size has to be big enough and small enough at the same time. It must be big enough for the Large Counts condition (np ≥ 10 and n(1−p) ≥ 10) or the CLT. It must be small enough relative to the population (n ≤ 10% of N) so that sampling without replacement still acts like independent draws. When n is 10% of N or less, you can ignore the finite population correction.

Is Sample Size on the AP® Statistics exam?

In multiple choice, sample size usually shows up as a "what happens if" question or a calculation. A typical one asks, "A population has σ = 8. For which n is the standard error of the mean 2?" (Answer: 8/√n = 2, so n = 16.) Other common ones ask for the minimum sample size to reach a target standard error or margin of error. Some ask when the finite population correction can be ignored (n ≤ 10% of the population), or how big n must be before a skewed population gives an approximately normal x̄.

In free response, n sits inside the conditions and the interpretation. The 2021 FRQ Q5 used counts from three independent samples of teens, so you have to work with the sample size of each group. The 2023 FRQ Q4 had a study with only nineteen patients. Small samples like that are a signal to check the normality conditions carefully, because the CLT can't rescue you. Released FRQs such as 2021 Q4 and 2024 Q1 also show you how n feeds into conditions, standard errors, and conclusions. Expect to explain in context how changing n would affect an interval's width or a test's power.

Sample Size vs Number of samples

Sample size (n) is how many observations are in ONE sample. The number of samples is how many separate samples you draw, as in a simulation that builds a sampling distribution from 1,000 samples of size 25. Increasing n makes the sampling distribution narrower and more normal. Increasing the number of samples just gives you a clearer picture of that same distribution without changing its spread. Its standard deviation is σ/√n, and the number of samples never appears in that formula.

Key things to remember about Sample Size

• Sample size (n) is the number of observations in a single sample, not the population size and not the number of samples drawn.

• The standard deviation of a sampling distribution shrinks by a factor of √n, so quadrupling the sample size cuts the standard error in half.

• A larger sample size gives a narrower confidence interval at the same confidence level and increases the power of a significance test.

• For skewed populations, the Central Limit Theorem needs a large enough n (often n ≥ 30) before the sampling distribution of x̄ is approximately normal.

• The sample should be no more than 10% of the population so that observations can be treated as independent when sampling without replacement.

• A bigger sample reduces variability but cannot fix bias, so a large poorly designed sample is still a bad sample.

Frequently asked questions about Sample Size

What is sample size in AP Statistics?

Sample size, written n, is the number of individuals or observations in a sample. It controls how much a statistic like x̄ or p̂ varies from sample to sample, through formulas like σ/√n.

Does a bigger sample size reduce bias?

No. A larger n reduces variability, but it does nothing about bias from a flawed method like a voluntary response survey or undercoverage. A huge biased sample just gives you a very precise wrong answer.

How does sample size affect margin of error?

Margin of error shrinks as n grows, but only by a factor of √n. To cut the margin of error in half, you need four times the sample size. For example, going from n = 100 to n = 400 halves it.

What is the difference between sample size and number of samples?

Sample size is how many observations are in one sample. Number of samples is how many separate samples you take, often in a simulation. Only sample size changes the spread of the sampling distribution, which is σ/√n.

How do you find the minimum sample size for a confidence interval?

Set the margin of error formula equal to your target and solve for n. For a proportion, use n ≥ (z*/ME)² · p(1−p), with p = 0.5 if you have no prior estimate. Always round up to the next whole number.