Skip to content
AP Statistics · Inference for Categorical Data: Proportions

3.6 p values

3.6 p values appears in 5 reviewed AP Statistics questions and sits in Inference for Categorical Data: Proportions. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How 3.6 p values is tested

Example 1Medium
A principal at a high school with 1,200 students believes that fewer than 40% of the students eat breakfast at home on school days. She selects a random sample of 30 students, and 9 of them (p-hat = 0.30) say they eat breakfast at home on school days. She plans to test H0: p = 0.40 versus Ha: p < 0.40 and find the p-value from a normal model. Which statement correctly assesses the conditions for this test?
  1. AThe 10% condition is not met because 30 students is less than 10% of the 1,200 students, so the responses are not independent.
  2. BAll conditions are met: the sample is random, 30 is under 10% of 1,200, and expected counts 12 and 18 are at least 10.Correct
  3. CThe Random condition is not met because a sample of only 30 students is too small to represent a school of 1,200 students.
  4. DThe Large Counts condition is not met because only 9 students in the sample said yes, which is fewer than 10, so the normal-based p-value is not reliable.

Why B is right

For a significance test about a proportion, the Large Counts condition is checked with the null value p0 = 0.40, because the p-value is computed assuming H0 is true: 30(0.40) = 12 and 30(0.60) = 18, and both are at least 10. The sample is random, and n = 30 is less than 10% of the population (0.10 x 1,200 = 120), so treating the observations as independent is reasonable. Because all conditions are met, the normal-based p-value can be trusted: SE = sqrt(0.40(0.60)/30) ≈ 0.0894, z = (0.30 - 0.40)/0.0894 ≈ -1.12, and the p-value ≈ 0.13.
Why the other 3 choices miss
A
Gets the 10% condition backwards. The sample must be no more than 10% of the population, and 30 < 120, so the condition is satisfied, not violated.
C
Mixes up sample size with random selection. The Random condition is about how the sample was chosen. A random sample of 30 meets it no matter how large the school is.
D
Uses the observed count (n times p-hat = 9) to check Large Counts. For a test, the check uses the hypothesized p0, which gives expected counts 12 and 18, so the condition is met.
Example 2Medium
A school district's nutrition coordinator thinks that more than 30 percent of the district's students skip breakfast on school days. To investigate, she selects a random sample of 400 students from the district, and 136 of them (p-hat = 0.34) report skipping breakfast on school days. She tests H0: p = 0.30 against Ha: p > 0.30, where p is the proportion of all students in the district who skip breakfast on school days. The test produces a p-value of about 0.040. Which of the following is a correct interpretation of this p-value?
  1. AThere is about a 0.960 probability that more than 30% of all students in the district skip breakfast on school days.
  2. BAssuming 30% of all students in the district skip breakfast, there is about a 0.040 probability of getting a sample proportion of 0.34 or greater in a random sample of 400 students.Correct
  3. CThere is about a 0.040 probability that exactly 30% of all students in the district skip breakfast on school days.
  4. DAssuming more than 30% of all students in the district skip breakfast, there is about a 0.040 probability of getting a sample proportion of 0.34 or greater.

Why B is right

A p-value is the probability of getting a sample result at least as extreme as the one observed, in the direction of Ha, calculated on the assumption that the null hypothesis is true. Here SE = sqrt(0.30(0.70)/400) = sqrt(0.000525) ≈ 0.0229, so z = (0.34 - 0.30)/0.0229 ≈ 1.75, and P(Z ≥ 1.75) ≈ 0.040. Choice B states this correctly: it assumes p = 0.30 and gives the probability of a sample proportion of 0.34 or more in a random sample of 400.
Why the other 3 choices miss
A
Takes 1 minus the p-value to be the probability that Ha is true. Neither 0.040 nor 0.960 is a probability that a hypothesis is true or false.
C
Treats the p-value as the probability that the null hypothesis is true. The p-value is calculated assuming H0 is true, so it cannot be the probability that H0 is true.
D
Assumes the alternative hypothesis instead of the null hypothesis. A p-value is always calculated under H0 (p = 0.30), not under Ha.

Studied alongside

Other concepts from Inference for Categorical Data: Proportions.