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

P values

P values appears in 7 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 P values is tested

Example 1Easier
A town library's director believes that more than 40% of its patrons now check out books at the self-service kiosks. To investigate, she selects a random sample of 200 patrons and finds that 92 of them, a sample proportion of 0.46, used the kiosks. She tests H0: p = 0.40 against Ha: p > 0.40, where p is the proportion of all the library's patrons who use the kiosks, and gets a p-value of about 0.042. Which of the following is the correct interpretation of this p-value?
  1. AIf p = 0.40, a random sample of 200 patrons has about a 0.042 chance of a sample proportion of 0.46 or more.Correct
  2. BThere is about a 0.042 probability that exactly 40% of all of the library's patrons use the self-checkout kiosks.
  3. CIf more than 40% of all patrons use self-checkout, the probability of getting a sample proportion of 0.46 is about 0.042.
  4. DThe probability of rejecting the null hypothesis that 40% of patrons use self-checkout when it is actually true is about 0.042.

Why A is right

A p-value is the probability, computed assuming the null hypothesis is true, of getting a sample statistic at least as extreme as the one observed, in the direction of the alternative. Here the null value is p = 0.40, the standard deviation of p-hat is sqrt(0.40*0.60/200) = 0.0346, so z = (0.46 - 0.40)/0.0346 = 1.73 and P(Z >= 1.73) = 0.042. Choice A states all three required parts: the null (p = 0.40) is assumed true, the result is "0.46 or more," and the sample size is 200.
Why the other 3 choices miss
B
Treats the p-value as the probability that the null hypothesis is true. A p-value is computed by assuming H0 is true, so it cannot be the probability that H0 is true.
C
Assumes the alternative instead of the null and leaves out the 'or more extreme' tail. The p-value is calculated under p = 0.40 and covers every sample proportion of 0.46 or higher, not only 0.46 itself.
D
Mixes up the p-value with the significance level (the Type I error probability), alpha. Alpha is chosen before the data are collected. The p-value comes from the observed sample.

Studied alongside

Other concepts from Inference for Categorical Data: Proportions.