Z-statistic in AP Statistics

The z-statistic is the standardized test statistic for a population proportion, z = (p̂ - p₀)/√(p₀(1-p₀)/n), which measures how many standard errors the sample proportion p̂ falls from the null hypothesis value p₀, assuming the null hypothesis is true.

Verified for the 2027 AP Statistics examLast updated June 2026

What is the z-statistic?

The z-statistic answers one question. If the null hypothesis is true, how surprising is my sample? You compute it as z = (p̂ - p₀)/√(p₀(1-p₀)/n), where p̂ is your sample proportion, p₀ is the value the null hypothesis claims, and n is the sample size. The result tells you how many standard errors your sample proportion sits away from the hypothesized proportion. A z of 0.4 means your sample landed basically where the null predicted. A z of 2.15 means it landed more than two standard errors out, which is unusual if the null is really true.

Under the CED, the general pattern is test statistic = (sample statistic - null value of the parameter) / (standard deviation of the statistic). For proportions, that standard deviation uses p₀, not p̂, because the whole calculation assumes the null hypothesis is true. The z-statistic only works as a shortcut to a theoretical normal distribution (the null distribution) when the conditions for normality are met. Otherwise, you'd compare your result to a randomization distribution instead.

Why the z-statistic matters in AP® Statistics

The z-statistic lives in Topic 6.5 (Interpreting p-Values) in Unit 6, Inference for Categorical Data: Proportions. Learning objective 6.5.A asks you to calculate an appropriate test statistic and p-value for a population proportion, and the z-statistic is exactly that calculation. Learning objective 6.5.B then asks you to interpret the p-value, which is the proportion of the null distribution as extreme or more extreme than your observed z. In other words, the z-statistic is the bridge between your raw data and the p-value. Every one-proportion z-test on the exam runs through it, and the same standardize-then-compare logic repeats for two proportions, means, and slopes later in the course.

How the z-statistic connects across the course

Z-Scores and the Standard Normal Distribution (Unit 1)

A z-statistic is just a z-score for a statistic instead of an individual data point. In Unit 1 you standardized one person's height; here you standardize an entire sample's proportion. Same idea, bigger object.

Sampling Distribution of p̂ (Unit 5)

The denominator √(p₀(1-p₀)/n) is the standard deviation of the sampling distribution of p̂ from Unit 5, with p₀ plugged in because you're assuming the null is true. Unit 5 built the distribution; Unit 6 uses it to judge surprise.

Null Distribution (Unit 6)

The z-statistic gets compared to the null distribution, which is what the test statistic looks like when H₀ is true. When a normal model applies, that null distribution is the standard normal (z); when it doesn't, a randomization distribution does the same job.

Significance Test and the p-Value (Unit 6)

The z-statistic feeds directly into the p-value. For Hₐ: p > p₀ you take the area above z, for Hₐ: p < p₀ the area below, and for Hₐ: p ≠ p₀ you take both tails. The z is the location; the p-value is the area.

Is the z-statistic on the AP® Statistics exam?

On multiple choice, expect to identify which value in z = (p̂ - p₀)/√(p₀(1-p₀)/n) is p₀ versus p̂, recognize the z-statistic as the standardized comparison between a sample proportion and a hypothesized one, and reason about what a given z (like z = 2.15) means for the p-value. A classic stem changes p₀ while keeping the same sample data and asks what happens to z and the p-value. On free response, the one-proportion z-test is a recurring full significance test, like the 2021 FRQ on whether a $10 coupon increases repeat purchases and the 2024 FRQ about a fitness center surveying members. You need to state hypotheses, check conditions, compute the z-statistic showing the formula with p₀ in the denominator, find the p-value, and write a conclusion in context. Using p̂ instead of p₀ in the standard error is one of the most commonly penalized errors.

The z-statistic vs z-score (Unit 1)

A z-score standardizes a single data value against a population mean and standard deviation. A z-statistic standardizes a sample statistic (like p̂) against a null hypothesis value, using the standard deviation of the sampling distribution. The arithmetic looks identical, but a z-score describes one observation while a z-statistic tests a claim about a parameter. That's why the z-statistic comes with a p-value attached and a z-score doesn't.

Key things to remember about the z-statistic

  • The z-statistic for a population proportion is z = (p̂ - p₀)/√(p₀(1-p₀)/n), and it counts how many standard errors the sample proportion falls from the null value.

  • Always use p₀, the null hypothesis value, in the standard error denominator, because the entire test statistic is calculated assuming the null hypothesis is true.

  • The general form is test statistic = (sample statistic - null value of the parameter) / (standard deviation of the statistic), a pattern that repeats throughout inference.

  • The z-statistic only maps to the standard normal curve when conditions hold; otherwise the null distribution comes from a randomization distribution instead.

  • The p-value is the area of the null distribution as extreme or more extreme than your z, with one tail for > or < alternatives and two tails for ≠.

  • A larger absolute z-statistic means stronger evidence against the null hypothesis, because it corresponds to a smaller p-value.

Frequently asked questions about the z-statistic

What is the z-statistic in AP Stats?

It's the standardized test statistic for a one-proportion significance test, z = (p̂ - p₀)/√(p₀(1-p₀)/n). It measures how many standard errors your sample proportion p̂ falls from the hypothesized proportion p₀, assuming the null hypothesis is true.

Do you use p̂ or p₀ in the z-statistic formula?

Use p₀ in the standard error, so the denominator is √(p₀(1-p₀)/n). The test assumes the null hypothesis is true, so the null value sets the spread. Using p̂ here is a common error that loses credit on FRQs. (Confidence intervals are the case where p̂ goes in the standard error.)

What's the difference between a z-statistic and a z-score?

A z-score standardizes one data point using the population mean and standard deviation. A z-statistic standardizes a sample statistic like p̂ against a null hypothesis value using the sampling distribution's standard deviation. Same math, but only the z-statistic produces a p-value for testing a claim.

Does a big z-statistic mean you reject the null hypothesis?

Usually, yes. A larger absolute z means your sample is further from what the null predicts, which shrinks the p-value. For example, z = 2.15 with a two-sided alternative gives a p-value around 0.03, below the typical α = 0.05 cutoff. The actual decision still depends on your significance level and the direction of the alternative.

How does the z-statistic relate to the p-value?

The p-value is the proportion of the null distribution as extreme or more extreme than your observed z-statistic. For Hₐ: p > p₀ it's the area above z, for Hₐ: p < p₀ the area below z, and for Hₐ: p ≠ p₀ it's both tails combined. That's exactly what learning objective 6.5.B asks you to interpret.