Null Hypothesis in AP Statistics

In AP Statistics, the null hypothesis (H₀) is a claim about a population parameter that states no effect, no difference, or no association, written with an equals sign (like H₀: p = 0.4). A significance test assumes H₀ is true and asks whether the sample data are unusual enough to reject it.

What is Null Hypothesis?

The null hypothesis, written H₀, is the "nothing interesting is going on" claim. It says a population parameter equals some specific value, or that two groups don't differ, or that two categorical variables aren't associated. Examples are H₀: μ = 200, H₀: p = 0.4, H₀: p₁ = p₂, and H₀: μ_diff = 0. For a chi-square test, it's stated in words, like "there is no association between gender and preference."

Think of H₀ as the defendant in a trial who is presumed innocent. You don't try to prove it. You assume it's true, then check whether your data would be really surprising in a world where it's true. That's exactly what a p-value measures. If the data are surprising enough (small p-value), you reject H₀ in favor of the alternative hypothesis. If not, you fail to reject H₀. You never "accept" or "prove" it, because not finding evidence against something isn't the same as showing it's true. H₀ is always about a population parameter (μ, p), never a sample statistic (x̄, p̂).

Why Null Hypothesis matters in AP® Statistics

The null hypothesis is the starting point of every significance test in the course. It first shows up when you set up a test for a population proportion (Topic 3.5). Then it becomes the assumption behind every p-value calculation (Topic 3.6) and every conclusion you write (Topic 3.7). The same structure repeats for two-proportion tests (3.12-3.13), chi-square tests for homogeneity or independence (3.14-3.15), and tests for means, paired differences, and differences in means (4.4-4.5, 4.9-4.10). It also defines what an error even is. A Type I error means rejecting H₀ when it was actually true (Topic 3.8). If you can't state H₀ correctly, the rest of the test falls apart. That includes the conditions, the test statistic, the p-value, and the conclusion in context.

How Null Hypothesis connects across the course

Alternative Hypothesis (Units 3-4)

H₀ and Hₐ are a matched pair about the same parameter. H₀ gets the equals sign, and Hₐ gets the <, >, or ≠ that matches the research question. The direction of Hₐ tells you which tail of the distribution to shade when you find the p-value, but the calculation itself always assumes H₀ is true.

p-value (Unit 3)

A p-value only makes sense in a world where H₀ is true. It's the probability of getting a statistic at least as extreme as yours if H₀ is true. So every correct p-value interpretation has to include the phrase "assuming the null hypothesis is true."

Type I Error (Unit 3)

A Type I error is a false alarm. You reject H₀ even though it was actually true. The significance level α is the probability of making this mistake, which is why lowering α makes it harder to reject H₀.

Chi-Square Statistic (Unit 3)

In chi-square tests, H₀ switches from a single number to a statement of "no association" or "same distribution." The expected counts are what you'd see if H₀ were true, and the chi-square statistic measures how far your observed counts stray from that H₀ world.

Is Null Hypothesis on the AP® Statistics exam?

Multiple choice. Questions hand you H₀ and Hₐ and ask you to interpret a p-value. For example, one gives H₀: μ = 200, Hₐ: μ > 200, and p = 0.03, and the correct answer must say "assuming μ = 200." Another type gives H₀: μ = 85 with Hₐ: μ < 85 and a test statistic of z = -1.28, and you find the p-value in the correct tail. Misconception questions are common too, such as a researcher claiming a p-value of 0.002 "proves" Hₐ. The fix is that small p-values give convincing evidence against H₀. They don't prove anything.

Free response. Inference FRQs ask whether the data "provide convincing statistical evidence." That's your cue to write hypotheses. Examples include 2017 Q5 (comparing age at diagnosis for men and women with schizophrenia) and 2018 Q6 (testing against a reported mean systolic blood pressure of 122). The term also appears in 2018 Q4 and 2019 Q4. For full credit you need to:

• Define the parameter in context ("μ = the true mean systolic blood pressure of...").
• Write H₀ with an equals sign using parameter notation.
• Compare the p-value to α.
• Conclude in terms of Hₐ in context ("there is/is not convincing evidence that...").

Null Hypothesis vs Alternative Hypothesis

The null hypothesis is the default "no effect" claim, and it always contains equality (=). The alternative hypothesis is what the researcher suspects or wants evidence for, and it uses <, >, or ≠. A quick check is that the claim you're trying to find evidence for goes in Hₐ, never H₀. You assume H₀ to do the math, but your conclusion is phrased as evidence for Hₐ.

Key things to remember about Null Hypothesis

• The null hypothesis (H₀) states no effect, no difference, or no association, and for parameter tests it always uses an equals sign.

• H₀ is always written about a population parameter like μ or p, never a sample statistic like x̄ or p̂.

• Every p-value is calculated assuming H₀ is true, so your interpretation should say so explicitly.

• You either reject H₀ or fail to reject H₀, and you never say you accepted or proved it.

• A Type I error happens when you reject a null hypothesis that was actually true, and its probability equals α.

• For chi-square tests, H₀ is written in words as "no association" (independence) or "the distributions are the same" (homogeneity).

Frequently asked questions about Null Hypothesis

What is the null hypothesis in AP Stats?

The null hypothesis (H₀) is the claim that a population parameter equals a specific value or that there's no difference or association, such as H₀: p = 0.4 or H₀: μ₁ = μ₂. A significance test assumes it's true and checks whether the sample data are unusual enough to reject it.

Does a large p-value prove the null hypothesis is true?

No. A large p-value only means you don't have convincing evidence against H₀, so you fail to reject it. The true parameter could still differ from the null value, and your sample just wasn't strong enough to show it.

How is the null hypothesis different from the alternative hypothesis?

H₀ is the no-change claim with an equals sign, while Hₐ is the claim you're looking for evidence of, using <, >, or ≠. For example, in a test of whether a new teaching method lowers scores, H₀: μ = 85 and Hₐ: μ < 85.

Can the null hypothesis use x̄ or p̂ instead of μ or p?

No. Hypotheses are always about population parameters, so you write μ or p, not x̄ or p̂. Writing a sample statistic in H₀ is a common way to lose credit on FRQs.

What is the null hypothesis for a chi-square test?

For a chi-square test for independence, H₀ says there is no association between the two categorical variables in the population. For a test for homogeneity, H₀ says the distribution of the categorical variable is the same across all populations or treatments.