A causal relationship exists when changing one variable directly produces a change in another. In AP Statistics, you can only conclude causation from a well-designed randomized experiment, because random assignment balances out confounding variables that observational studies cannot rule out.
A causal relationship means one variable actually makes another variable change. If you increase the dose of a drug and blood pressure drops because of the drug, that's causation. This is different from two variables simply moving together (correlation), which can happen for lots of reasons that have nothing to do with cause and effect.
The core AP Stats rule is this: the scope of your conclusion depends on how the data were collected. In an observational study, researchers just watch what happens, so a lurking or confounding variable could be driving the association. Ice cream sales and drowning deaths rise together, but hot weather causes both. In a randomized experiment, treatments are randomly assigned to experimental units, which spreads confounding variables roughly evenly across groups. That's why random assignment is the one thing that earns you the right to say "X causes Y." No random assignment, no causal claim. You can only say the variables are associated.
Causal relationships sit at the intersection of two big chunks of the course. In Unit 2 (Exploring Two-Variable Data), you learn that correlation measures the strength of a linear association, and the CED expects you to know that correlation does not imply causation. In Unit 3 (Collecting Data), you learn why, and what fixes it. The distinction between observational studies and experiments exists almost entirely to answer the question "when can we claim causation?" Random assignment of treatments is the mechanism that makes causal conclusions valid.
This idea also feeds the exam's scope-of-inference framework, which shows up constantly. Random selection lets you generalize to a population; random assignment lets you claim causation. Mixing those two up is one of the most common ways to lose points on study-design questions.
Keep studying AP Statistics Unit 3
Correlation (Unit 2)
Correlation tells you two quantitative variables move together; causation tells you one drives the other. A strong r value, even r = 0.95, never proves cause and effect on its own. The classic exam trap gives you a strong correlation from observational data and asks if a causal conclusion is justified. The answer is no.
Confounding Variable (Unit 3)
A confounding variable is the main reason observational studies can't support causal claims. It's a third variable tangled up with both the explanatory and response variables, so you can't tell which one is actually doing the causing. Naming a plausible confounder is a standard FRQ move.
Experimental Design (Unit 3)
Random assignment is the magic ingredient. By randomly assigning treatments, an experiment balances confounding variables (known and unknown) across groups, so a difference in outcomes can be attributed to the treatment itself. This is the only data-collection method that licenses a cause-and-effect conclusion.
Observational Study (Unit 3)
In an observational study, researchers measure variables without imposing treatments. You can detect associations, but because nothing controls for confounding, the strongest conclusion you can draw is that the variables are associated, not that one causes the other.
Causal relationships show up most often as a judgment call about study design. A multiple-choice stem will describe a study, report an association, and ask which conclusion is appropriate. The right answer hinges on one question: were treatments randomly assigned? FRQs do the same thing in written form. A common task gives you an observational result and asks whether a causal conclusion is justified (say no, and explain that a confounding variable could explain the association, ideally naming a plausible one), or asks you to design an experiment that would allow a causal conclusion, where you need to describe random assignment explicitly. Later in the course, inference FRQs about experiments often end by asking what conclusion the design supports, and "random assignment allows a cause-and-effect conclusion" is exactly the sentence graders look for. No released FRQ needs the phrase "causal relationship" verbatim; what's tested is whether you apply the logic correctly.
Correlation is a number (r) describing how strongly two quantitative variables move together in a straight-line pattern. A causal relationship is a claim that one variable produces changes in the other. Correlation is evidence of association, but association can come from confounding, coincidence, or reverse causation. Only random assignment in an experiment turns an observed association into a defensible causal claim. Memorize the line: correlation does not imply causation.
A causal relationship means changing one variable directly produces a change in another variable, not just that the two move together.
Only a well-designed experiment with random assignment of treatments can justify a cause-and-effect conclusion.
Observational studies can only show association, because a confounding variable could explain the relationship.
Random assignment supports causal conclusions; random selection supports generalizing to a population. These are two different things and the exam tests both.
When an FRQ asks if a causal conclusion is appropriate for observational data, say no and name a plausible confounding variable.
It's a relationship where one variable directly affects another, so changing the explanatory variable produces a change in the response variable. In AP Stats, you can only conclude a causal relationship from a randomized experiment.
No. Even a correlation near r = 1 or r = -1 only shows a strong linear association. A confounding variable, coincidence, or reverse causation could explain the pattern, which is why "correlation does not imply causation" is a core exam rule.
Association means two variables tend to vary together; causation means one actually drives changes in the other. Observational studies can establish association, but only experiments with random assignment can establish causation.
Because researchers don't control who gets which treatment, a confounding variable can influence both the explanatory and response variables. For example, hot weather drives up both ice cream sales and drownings, creating an association with no causation between them.
No. A random sample lets you generalize results to the population, but without randomly assigned treatments you can't rule out confounding. Generalization and causation come from two separate design features.
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.