Skip to content
AP Statistics · Probability, Random Variables, and Probability Distributions

Simpsons paradox

Simpsons paradox appears in 6 reviewed AP Statistics questions and sits in Probability, Random Variables, and Probability Distributions. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Simpsons paradox is tested

Example 1Harder
An admissions office reports that the acceptance rate for male applicants is higher than for female applicants in every individual department at the university. However, the overall acceptance rate for female applicants across all departments is higher than for males. Which statement best explains this result?
  1. AFemale applicants applied more often to departments with higher acceptance ratesCorrect
  2. BThis is impossible; the data must contain an error
  3. CThe sample sizes within each department are too small to produce reliable acceptance rates
  4. DThe university is practicing gender discrimination that only appears when looking at departmental data

Why A is right

This is Simpson's paradox. Even though males have a higher acceptance rate within each department, if females disproportionately apply to departments that accept a larger proportion of all applicants, the overall female acceptance rate can exceed the male rate. The lurking variable is the department to which applicants apply.
Why the other 3 choices miss
B
Incorrect because Simpson's paradox is a well-documented statistical phenomenon that can occur with real data; it does not indicate an error.
C
Incorrect because the paradox arises from the structure of the data (different application patterns), not from small sample sizes.
D
Incorrect because the departmental data show males being accepted at higher rates within each department, so the disaggregated data do not support a claim of discrimination against males.

Studied alongside

Other concepts from Probability, Random Variables, and Probability Distributions.