Experiment in AP Statistics

In AP Statistics, an experiment is a study where researchers deliberately impose treatments on experimental units, ideally through random assignment. That random assignment is what lets you draw cause-and-effect conclusions, which an observational study can't support.

What is Experiment?

An experiment is a study where the researcher does something to the subjects. They pick the treatments, decide who gets which one, and then measure the response. That's the whole difference from an observational study, where researchers just watch and record what people already chose to do.

The pieces have names you need to know. The experimental units are whatever receives the treatment (people, plants, plots of trees). The explanatory (independent) variable is what you change. The response variable is what you measure. A well-designed experiment uses random assignment to put units into groups. It often includes a control group for comparison, and it repeats the treatments on enough units to see a real pattern. Think of random assignment as a coin flip that keeps the researcher's bias, and the subjects' own choices, out of the groups. If the groups start out roughly equal, a difference at the end can be blamed on the treatment.

Why Experiment matters in AP® Statistics

Experiments live in Topic 1.10, The Investigative Question Revisited and Data Collection. The big idea there is that how you collect the data controls which conclusions you're allowed to make. A good statistical question isn't finished until you pick a design that can actually answer it. If the question is "does X cause Y?", only a randomized experiment gets you there. This idea comes back every time you write a conclusion in context. The exam repeatedly asks whether a cause-and-effect claim is justified, and the answer almost always depends on whether treatments were randomly assigned.

How Experiment connects across the course

Random Assignment (Unit 1)

Random assignment is the engine of an experiment. Without it you have a comparison, but not a fair one. Randomly splitting units into groups spreads out hidden differences like fitness, motivation, or soil quality. That way the treatment is the only systematic difference left.

Causal Relationships vs. Association (Unit 1)

Observational data can show an association, like coffee drinkers living longer. It can't rule out a confounding variable such as exercise. A randomized experiment is the tool that turns "these go together" into "this causes that."

Control Group and Comparison (Unit 1)

A treatment result means nothing without a baseline. The control group (no treatment, a placebo, or the standard treatment) tells you what would have happened anyway. Then you can see how much of the change the treatment actually produced.

Random Sampling vs. Random Assignment and Scope of Inference (Unit 1)

These two randomizations answer different questions. Random sampling lets you generalize to a population. Random assignment lets you claim causation. A study can have one, both, or neither, and your conclusion has to match exactly what the design earned.

Is Experiment on the AP® Statistics exam?

Experiments show up in both multiple choice and free response, and the exam loves to test design flaws.

Multiple choice: Watch for setups where subjects choose their own treatment. One example is a gardener letting customers pick which fertilizer to use. Another is a teacher letting students choose a morning or afternoon section. The flaw is that there's no random assignment, so a confounding variable could explain the result. A causal conclusion isn't justified. You'll also see observational studies dressed up as evidence of causation (coffee and longevity). There, you have to point out that an uncontrolled variable like exercise could be responsible.

Free response: Released FRQs have built entire questions around experiments.

• 2017 Q6 assigned two men and two women to treatment and control groups.
• 2018 Q4 used a knee-surgery (ACL) recovery context.
• 2019 Q2 involved four concentrations of a fungus mixture used against tree-destroying insects.
• 2021 Q2 described a random sample of 100 adults in a walking-and-cholesterol study.

On these questions you might identify treatments, units, and the response variable. You might describe a random assignment method clearly enough that someone else could carry it out, or explain why a control group matters. You may also have to state the scope of inference, meaning whether results generalize to a population, show causation, or both.

Experiment vs Observational Study

Ask one question. Did the researcher impose the treatment? If yes, it's an experiment. If the researcher only observed groups that formed on their own (coffee drinkers vs. non-drinkers, students who picked a section), it's an observational study. Comparing two groups doesn't make a study an experiment. Only experiments with random assignment can support cause-and-effect conclusions.

Key things to remember about Experiment

• An experiment is a study where researchers deliberately impose treatments on experimental units and then measure a response.

• Random assignment is what allows an experiment to support a cause-and-effect conclusion, because it balances out confounding variables between groups.

• If subjects choose their own treatment, the study loses the protection of random assignment, and a causal conclusion is not justified.

• A control group gives you a baseline, so you can tell how much of the response was actually caused by the treatment.

• Random sampling lets you generalize to a population, while random assignment lets you claim causation, and your conclusion should only claim what the design supports.

• On FRQs, describe random assignment specifically enough that another person could carry it out exactly as you wrote it.

Frequently asked questions about Experiment

What is an experiment in AP Statistics?

An experiment is a study where researchers deliberately assign treatments to experimental units and measure the response. When the treatments are randomly assigned, the results can support a cause-and-effect conclusion.

Is a study an experiment if it compares two groups?

Not necessarily. If people chose their own group, like customers picking which fertilizer to use, it's an observational study even though two groups are compared. It's only an experiment when the researcher imposes the treatments.

What's the difference between an experiment and an observational study?

In an experiment, the researcher imposes treatments, ideally through random assignment. In an observational study, the researcher just records what subjects already do. Only a randomized experiment can show causation, which is why a study linking coffee to longer life can't prove coffee is the cause.

Can you generalize experiment results to a whole population?

Only if the experimental units were randomly sampled from that population. In the 2021 FRQ Q2 setup, researchers took a random sample of 100 adults. That kind of design lets you generalize to the target population, and random assignment on top of it would also support causation.

Do experiments always need a control group?

No, but they always need a comparison. Some experiments compare several treatments against each other, like the four fungus concentrations in 2019 FRQ Q2. A control group (no treatment or a placebo) is common because it gives a clear baseline.