Causal Relationships in AP Statistics

In AP Statistics, a causal relationship exists when changes in an explanatory variable directly cause changes in a response variable. You can only conclude causation from a well-designed experiment that randomly assigns treatments to experimental units. Observational studies can show association, not causation.

What is Causal Relationships?

A causal relationship means one variable actually makes the other change. It isn't just that the two variables move together. If you give one group a new fertilizer and another group none, and the fertilized plants grow taller because of the fertilizer, that's causation.

In AP Stats, the big question is never "does causation exist?" It's "does this study let you claim it?" That depends on how the data were collected (Topic 1.10). In an experiment, the researcher assigns treatments to experimental units. In an observational study, the researcher just records what's already happening. When treatments are randomly assigned, the groups should be similar in every way except the treatment. So a difference in the response can be pinned on the treatment. An observational study has no assignment. Some other variable (a confounding variable) could explain the difference, so the most you can claim is an association.

Why Causal Relationships matters in AP® Statistics

Causal relationships sit in Unit 1, Topic 1.10 (The Investigative Question Revisited and Data Collection). They tie directly to AP Stats 1.10.C (identify an experiment) and AP Stats 1.10.D (identify an observational study), because the study type decides whether a cause-and-effect conclusion is allowed. They also pair with AP Stats 1.10.E (justify generalizations). That objective covers the other half of the scope-of-inference question: random sampling tells you who you can generalize to. Random assignment tells you whether you can claim causation. Every inference procedure in Units 3-5 ends with a conclusion, and the wording of that conclusion depends on this distinction. Overclaiming causation is one of the most common ways to lose points.

How Causal Relationships connects across the course

Experiments and Random Assignment (Unit 1)

Random assignment is the ticket to a causal claim. It spreads out lurking variables evenly across treatment groups. Once that's done, the treatment is the only systematic difference left, so it's the only thing that can explain a difference in the response variable.

Random Sampling and Generalization (Unit 1)

Think of scope of inference as two separate switches. Random sampling flips the "generalize to the population" switch (AP Stats 1.10.E). Random assignment flips the "cause and effect" switch. A study can have one, both, or neither, and you need to say which.

Confidence Intervals and Tests for Means (Unit 4)

A statistically significant difference doesn't upgrade an observational study into an experiment. If a confidence interval for the difference in mean cholesterol between oat eaters and non-oat eaters excludes 0, you've found convincing evidence of a difference, not proof that oats cause it.

Correlation and Regression (Unit 5)

A strong correlation or a high r² tells you how well a line fits the data. It doesn't tell you why. Unless the data came from a randomized experiment, a slope describes an association, not the effect of changing x.

Is Causal Relationships on the AP® Statistics exam?

You'll rarely be asked to define causation. You'll be asked to judge whether a conclusion is justified. Typical MCQ stems describe a study and ask which conclusion is appropriate. The trap answer says one variable "causes" or "leads to" another when the study was observational. For example, a study comparing sleep duration for tea drinkers versus coffee drinkers is observational, so even a big gap in mean sleep can't show that tea causes more sleep. People who choose tea might differ in age, stress, or schedule.

In FRQs, causation shows up inside scope-of-inference questions and inference conclusions. No released FRQ has used the phrase "causal relationships" verbatim, but FRQs regularly ask whether a cause-and-effect conclusion can be drawn. To earn credit, you need to (1) identify whether treatments were randomly assigned, (2) state the correct conclusion (causation or only association), and (3) when it's observational, name a plausible confounding variable and explain how it could affect the response.

Causal Relationships vs Association (Correlation)

An association means two variables tend to change together. A causal relationship means one actually produces the change in the other. Every causal relationship creates an association, but plenty of associations aren't causal. Ice cream sales and drowning deaths rise together because both go up in summer, not because ice cream causes drowning. In AP Stats, the dividing line is study design. Observational studies can show association. Only randomized experiments can support causation.

Key things to remember about Causal Relationships

• A causal relationship means a change in the explanatory variable actually produces a change in the response variable.

• Only a well-designed experiment with random assignment of treatments lets you conclude cause and effect.

• Observational studies, including surveys, prospective studies, and retrospective studies, can show an association but not causation, because confounding variables can't be ruled out.

• Random sampling decides who you can generalize to, and random assignment decides whether you can claim causation, so treat them as two separate questions.

• A statistically significant result or a strong correlation does not prove causation if the data came from an observational study.

• When you reject a causal claim on the exam, name a specific confounding variable and explain how it could be linked to both variables.

Frequently asked questions about Causal Relationships

What is a causal relationship in AP Statistics?

It's a relationship where changing the explanatory variable directly causes a change in the response variable. In AP Stats, you can only claim one when the data come from a randomized experiment, as covered in Topic 1.10.

Can an observational study prove causation?

No. Because the researcher doesn't assign treatments, a confounding variable could explain any difference. An observational study, like one comparing cholesterol in daily oat eaters versus non-oat eaters, can only show an association.

What's the difference between correlation and causation in AP Stats?

Correlation (association) means two variables move together, while causation means one actually changes the other. A strong r in Unit 5 or a significant test in Units 3-4 shows association. Only random assignment in an experiment supports causation.

If a confidence interval doesn't contain 0, does that mean the treatment caused the difference?

Only if the study was a randomized experiment. An interval that excludes 0 gives convincing evidence of a real difference, but for an observational study you still can't say the explanatory variable caused it.

Does random sampling let you conclude cause and effect?

No, random sampling lets you generalize to the population (AP Stats 1.10.E), but causation requires random assignment of treatments. A random sample of tea and coffee drinkers still can't show that tea causes longer sleep.