Independence in AP Statistics

In AP Statistics, two events are independent if knowing one occurred doesn't change the probability of the other, so P(A|B) = P(A). For inference, the independence condition means observations don't influence each other, which you justify with random sampling or assignment and the 10% condition.

What is Independence?

Independence shows up in AP Stats in two connected ways.

First, in probability, events A and B are independent if learning that B happened tells you nothing new about A. In symbols, P(A|B) = P(A). The multiplication shortcut P(A and B) = P(A) · P(B) works only because of independence. Think of flipping a coin twice. The first flip has no memory, so it can't push the second flip around.

Second, in inference, independence is a condition you check before building a confidence interval or running a test. It means each observation in your data doesn't affect the others. You can't prove this directly, so you back it up two ways. You point to random sampling or random assignment, and when sampling without replacement you check the 10% condition (n is no more than 10% of the population). For two-sample procedures, you also need the two groups themselves to be independent of each other, and that's exactly why paired data gets its own method. Finally, the chi-square test for independence turns the probability idea into a hypothesis. It asks whether two categorical variables are associated or not.

Why Independence matters in AP® Statistics

Independence is the thread running through nearly every inference topic. It sits underneath the sampling distributions in Topic 2.12 (Sampling Distributions and the Central Limit Theorem), because the standard deviation formulas for p̂ and x̄ assume independent observations. From there it becomes a required condition in every "Setting Up" and "Constructing" topic: one- and two-proportion intervals and tests (3.3, 3.5, 3.10, 3.12), and one-sample, paired, and two-sample mean procedures (4.2, 4.4, 4.7, 4.9). In Topic 3.14 it stops being just a condition and becomes the hypothesis itself in the chi-square test for independence. On the exam, skipping or botching the independence check costs you points on inference FRQs. Graders look for the condition named and tied to the context.

How Independence connects across the course

Sampling Distributions and the CLT (Unit 2)

The formula SD = √(p(1−p)/n) only works if observations are independent. The 10% condition is basically your permission slip to use that formula when you sample without replacement. Without independence, the spread of your sampling distribution is wrong, and everything built on it is wrong too.

Independence Condition in Proportion Inference (Unit 3)

Every one- and two-proportion interval or test opens with the same checklist: random, 10%, and large counts. For two proportions, you need independence within each sample and between the two samples. That's why two separate random samples (like residents of two regions) qualify.

Chi-Square Test for Independence (Unit 3)

Here independence flips from a condition into the null hypothesis. H₀ says two categorical variables have no association, meaning knowing one tells you nothing about the other. It's the probability rule P(A|B) = P(A), tested on a whole two-way table at once.

Independent Groups vs. Paired Data in Mean Inference (Unit 4)

Two-sample t procedures (4.7, 4.9) need independent groups. Paired data (4.2, 4.4) comes from matched units, so the two measurements are dependent. You handle that by taking differences first and running a one-sample procedure on them. Picking the right method starts with asking "are these groups independent?"

Is Independence on the AP® Statistics exam?

Independence is tested constantly, though usually as a step inside a bigger question rather than as a standalone vocabulary item.

Multiple choice: Practice questions comparing two groups (side effects for medication A vs. B, policy support in two regions, exercise rates in City A vs. City B, male vs. female students working part-time) ask which conditions are met or which procedure fits. You need to recognize that two separate random samples give independent groups, and you need to verify the 10% condition for each sample.

Free response: Released FRQs set up exactly these comparisons. 2017 Q5 compares age at diagnosis for a random sample of men and women. 2019 Q3 summarizes medication-adherence responses (never, sometimes, always) for men and women, which is the kind of two-way categorical setup where association and independence come up. 2021 Q4 describes a company testing a $10-off offer, an experimental design where random assignment does the independence work. On inference FRQs, you have to:

• Name the condition (random sample or random assignment, plus 10% when sampling without replacement).
• Connect it to context ("the 200 residents are less than 10% of all residents of region 1").
• Choose the right procedure based on whether groups are independent or paired.
• In probability questions, check independence by comparing P(A|B) to P(A) with actual numbers.

Independence vs Mutually exclusive events

Mutually exclusive means A and B can't happen together, so P(A and B) = 0. Independent means one happening doesn't change the chance of the other. These almost never overlap. If two events with nonzero probabilities are mutually exclusive, then knowing A happened tells you B definitely didn't, so they are actually dependent. Quick test: mutually exclusive is about overlap in the Venn diagram, and independence is about whether P(A|B) equals P(A).

Key things to remember about Independence

• Two events are independent when P(A|B) = P(A), meaning knowing one happened doesn't change the probability of the other.

• The multiplication rule P(A and B) = P(A) · P(B) only works for independent events, so check independence before you multiply.

• For inference, you justify independence with random sampling or random assignment and, when sampling without replacement, the 10% condition.

• Two-sample procedures need the two groups to be independent, while paired data is dependent and gets analyzed as one set of differences.

• In a chi-square test for independence, the null hypothesis is that two categorical variables have no association.

• Mutually exclusive events with nonzero probabilities are dependent, not independent, which is a classic multiple-choice trap.

Frequently asked questions about Independence

What is independence in AP Statistics?

Independence means one event's outcome doesn't affect another's probability, so P(A|B) = P(A). In inference, it's the condition that observations don't influence each other, which you support with randomization and the 10% condition.

Are mutually exclusive events independent?

No. If two events with nonzero probabilities are mutually exclusive, knowing one happened means the other definitely didn't, so they're dependent. Mutually exclusive means P(A and B) = 0, while independent means P(A and B) = P(A) · P(B).

Does the 10% condition prove independence?

Not by itself. The 10% condition only says that sampling without replacement is close enough to independent when n is at most 10% of the population. You still need random sampling or random assignment, so state both in your conditions check.

What's the difference between a chi-square test for independence and a test for homogeneity?

A test for independence uses one random sample and asks whether two categorical variables are associated. A test for homogeneity uses separate samples from two or more populations and asks whether one categorical variable has the same distribution in each. The calculations match, but the design, hypotheses, and conclusion wording differ.

How do I know if samples are independent or paired on the AP exam?

Ask whether each value in one group is naturally matched to a specific value in the other, like before-and-after measurements on the same person. If it is, the data are paired (dependent) and you use a one-sample procedure on the differences. If the groups are separate random samples, like 200 residents of one region and 250 of another, they're independent and you use a two-sample procedure.