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

Two proportion z test

Two proportion z test appears in 30 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 Two proportion z test is tested

Example 1Medium
A two-proportion z-test compares the proportion of smokers in two cities. City A: n=200, x=42 smokers. City B: n=300, x=54 smokers. What is the pooled sample proportion p̂_c?
  1. Ap̂_c = 0.225
  2. Bp̂_c = 0.210
  3. Cp̂_c = 0.192Correct
  4. Dp̂_c = 0.180

Why C is right

The pooled (combined) sample proportion is found by pooling all successes over all observations from both samples: p̂_c = (x₁ + x₂)/(n₁ + n₂) = (42 + 54)/(200 + 300) = 96/500 = 0.192. This corresponds to choice C.
Why the other 2 choices miss
A
Incorrect because 0.225 results from averaging the two proportions (0.21 + 0.18)/2 = 0.195, close but incorrect; or from miscounting the smokers.
B
Incorrect because 0.210 results from using only one city's proportion (42/200 = 0.21) rather than combining both cities' data.
Example 2Medium
A city planner wants to know whether support for a proposed protected bike lane is higher among residents of the North district than among residents of the South district. She selects a random sample of 220 North district residents and an independent random sample of 180 South district residents. Of these, 132 North residents and 90 South residents say they support the bike lane. Which of the following are the correct hypotheses for a significance test that answers the planner's question?
  1. AH0: p-hat_N - p-hat_S = 0 and Ha: p-hat_N - p-hat_S > 0, where p-hat_N and p-hat_S are the proportions of the sampled North and South residents who support the bike lane
  2. BH0: p_N - p_S = 0 and Ha: p_N - p_S > 0, where p_N and p_S are the proportions of all North and all South district residents who support the bike laneCorrect
  3. CH0: p_N - p_S = 0 and Ha: p_N - p_S ≠ 0, where p_N and p_S are the proportions of all North and all South residents who support the bike lane
  4. DH0: p_N - p_S = 0.10 and Ha: p_N - p_S > 0.10, where p_N and p_S are the proportions of all North and South residents who support the lane

Why B is right

Hypotheses are always statements about population parameters. Here the parameters are the true proportions of all North and all South district residents who support the bike lane. The null hypothesis says there is no difference (p_N - p_S = 0). Because the planner asks whether support is higher in the North, the alternative is one-sided: p_N - p_S > 0. The sample proportions (132/220 = 0.60 and 90/180 = 0.50) are used to compute the test statistic. They do not belong in the hypotheses.
Why the other 3 choices miss
A
States the hypotheses in terms of sample statistics (p-hat). We already know the sample proportions (0.60 and 0.50), so there is nothing to test about them. Hypotheses must be about population parameters.
C
Uses a two-sided alternative. The planner's question is directional (is support higher in the North?), so the alternative should be p_N - p_S > 0, not ≠ 0.
D
Puts the observed sample difference, 0.60 - 0.50 = 0.10, into the null hypothesis. The null value is chosen before looking at the data and should be 0 here, meaning no difference between the districts.

Studied alongside

Other concepts from Inference for Categorical Data: Proportions.