Skip to content
AP Statistics · Exploring One-Variable Data and Collecting Data

Randomized experiment

Randomized experiment appears in 6 reviewed AP Statistics questions and sits in Exploring One-Variable Data and Collecting Data. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Randomized experiment is tested

Example 1Medium
A researcher studying climate data finds that r = 0.88 between average summer temperature and annual wildfire area (in thousands of acres) across 30 regions. Which of the following would provide the strongest evidence that increased temperatures CAUSE increased wildfire area?
  1. AIncreasing the sample to 300 regions and showing the correlation is statistically significant.
  2. BFitting a regression model with r^2 > 0.90.
  3. CConducting a randomized experiment that randomly assigns temperatures to regions.Correct
  4. DFinding r = 0.99 instead of 0.88 in a follow-up study.

Why C is right

Causation can only be established through a randomized experiment where the explanatory variable (temperature) is randomly assigned. In practice, randomly assigning temperatures to regions is not feasible, but the principle stands: only random assignment of the treatment controls for confounding variables and allows causal inference. Observational studies (including correlations) can only show association.
Why the other 3 choices miss
A
Incorrect because increasing sample size and achieving statistical significance means the observed association is unlikely to be due to chance, but it does not eliminate confounding or establish causation. Statistical significance ≠ causation.
B
Incorrect because r^2 > 0.90 is a measure of explanatory power in the regression model. A higher r^2 still cannot establish causation — it only means more variation in the response is explained by the predictor.
D
Incorrect because a stronger correlation (r = 0.99) still only shows stronger association, not causation. No value of r, no matter how close to 1, can establish causation from observational data.

Studied alongside

Other concepts from Exploring One-Variable Data and Collecting Data.