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

One sample z test

One sample z test 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 One sample z test is tested

Example 1Easier
A school official claims that 30 percent of the students at a large high school buy lunch from the cafeteria on a typical day. A student council member thinks the proportion is different and plans to select a random sample of 150 students to test H0: p = 0.30 versus Ha: p ≠ 0.30, where p is the proportion of all students at the school who buy cafeteria lunch on a typical day. Before collecting any data, she computes 150(0.30) = 45 and 150(0.70) = 105. Which of the following is the best interpretation of what these calculations show?
  1. AAssuming p = 0.30, both expected counts are at least 10, so the sampling distribution of p-hat is approximately normal and a one-sample z-test is appropriate.Correct
  2. BExactly 45 of the 150 sampled students will buy cafeteria lunch, so p-hat will equal 0.30.
  3. CBecause both values are at least 10, the distribution of lunch purchases in the student population is approximately normal, so a one-sample z-test for p can be used.
  4. DThe sample is large enough to show that the proportion of students who buy cafeteria lunch is 0.30.

Why A is right

The calculations check the large counts condition for a significance test, which uses the null value: n*p0 = 150(0.30) = 45 and n(1 - p0) = 150(0.70) = 105. Both are at least 10, so if H0 is true, the sampling distribution of p-hat is approximately normal. That justifies using a one-sample z-test for a proportion. The calculations say nothing about whether H0 is true, what the sample will show, or the shape of the population.
Why the other 3 choices miss
B
Confuses the expected count under H0 with the observed count. Because of sampling variability, the actual number of lunch buyers in a random sample of 150 will vary around 45 and will rarely equal it exactly.
C
Confuses the sampling distribution of p-hat with the population distribution. Each student either buys lunch or does not, so the population is categorical (yes/no). The large counts condition is about the approximate normality of p-hat, not the population, so this reasoning for the z-test is wrong.
D
Treats a check of conditions as evidence about the parameter. Meeting the large counts condition only makes the test procedure valid, and no data have been collected yet; a test also never proves that H0 is true.

Studied alongside

Other concepts from Inference for Categorical Data: Proportions.