Difference of Two
Population Proportions
Learn how to carry out a significance test comparing two population proportions: state hypotheses, check conditions, compute the z-statistic, and draw conclusions.
When Do We Compare Proportions?
Many real-world questions involve comparing two groups: Does a new drug cure a higher proportion of patients than a placebo? Do men and women differ in their support for a policy? Each group has its own population proportion, \( p_1 \) and \( p_2 \).
Parameters vs. Statistics
We never observe the true population proportions directly. Instead, we collect random samples and compute sample proportions.
The Big Picture: Test Framework
Every two-proportion z-test follows four stages: state the null and alternative hypotheses, verify the conditions, compute the z-statistic and p-value, and make a decision in context.
Setting Up Hypotheses
The null hypothesis always assumes no difference. The alternative depends on the research question.
Checking Conditions
The Pooled Proportion \( \hat{p}_c \)
Under \( H_0 \), we assume \( p_1 = p_2 \). Our best single estimate of that common proportion pools both samples:
This pooled estimate is used in both the standard error formula and the Large Counts condition for a hypothesis test.
Computing the z-Statistic
P-value and Decision
Test vs. Confidence Interval
| Significance Test | Confidence Interval | |
|---|---|---|
| Purpose | Is there a difference? | How big is the difference? |
| Standard Error | Pooled \( \hat{p}_c \) | Unpooled \( \hat{p}_1, \hat{p}_2 \) |
| Output | z-statistic and p-value | Interval estimate |
| Large Counts | Uses \( \hat{p}_c \) | Uses \( \hat{p}_1 \) and \( \hat{p}_2 \) |
Step-by-Step Procedure
Example 1: Vaccine Trial
A researcher tests whether a new vaccine reduces infections. In the treatment group, 40 of 200 got infected. In the control group, 60 of 200 got infected. Test at \( \alpha = 0.05 \).
\( \hat{p}_1 = 0.20 \)
\( \hat{p}_2 = 0.30 \)
Example 1: Calculations
Example 2: Voter Preference
In a random sample, 120 of 400 women and 100 of 400 men support a new policy. Is there a significant difference at \( \alpha = 0.05 \)?
Example 3: Online vs. In-Store
A retailer randomly surveys customers. Of 150 online shoppers, 45 returned an item. Of 250 in-store shoppers, 50 returned an item. Test whether online return rates are higher (\( \alpha = 0.01 \)).
Common Mistakes
Worked Example
A researcher randomly assigns 250 patients to receive a new drug and 250 patients to receive a placebo. In the drug group, 180 out of 250 show symptom improvement. In the placebo group, 155 out of 250 show improvement. At α = 0.05, is there convincing evidence that the drug produces a higher improvement rate than the placebo? Step 1 — State Hypotheses Let p₁ = the true proportion who improve with the drug and p₂ = the true proportion who improve with the placebo. H₀: p₁ − p₂ = 0 (no difference). Hₐ: p₁ − p₂ > 0 (the drug is better). This is a one-sided test (right-tailed). Step 2 — Check Conditions Random: Patients were randomly assigned to treatment groups. ✓ Independent: Random assignment supports the treatment comparison. The z-test also assumes independent individual responses and no material clustering or interference; these conditions must be justified by the study design. Large Counts: We need p̂_c first (computed in Step 3): p̂_c = 0.67. Then n₁p̂_c = 250 × 0.67 = 167.5 ≥ 10, n₁(1−p̂_c) = 82.5 ≥ 10, n₂p̂_c = 167.5 ≥ 10, n₂(1−p̂_c) = 82.5 ≥ 10. ✓ The large-count condition is met; the analysis is conditional on the independence assumptions above. Step 3 — Compute Pooled Proportion p̂₁ = 180/250 = 0.72. p̂₂ = 155/250 = 0.62. The pooled proportion is: p̂_c = (180 + 155) / (250 + 250) = 335 / 500 = 0.67 Step 4 — Compute Standard Error SE = √[ 0.67 × 0.33 × (1/250 + 1/250) ] = √[ 0.2211 × 0.008 ] = √0.001769 SE ≈ 0.04206 Step 5 — Compute Test Statistic z = (p̂₁ − p̂₂) / SE = (0.72 − 0.62) / 0.04206 = 0.10 / 0.04206 z ≈ 2.377 Step 6 — Find P-Value Since Hₐ is one-sided (>), the p-value = P(Z ≥ 2.377). Using the standard normal table or calculator: p-value ≈ 0.0087. p-value ≈ 0.0087 Step 7 — State Conclusion in Context Conditional on those study-design assumptions, because the p-value (0.0087) is less than α = 0.05, we reject H₀. There is convincing evidence that the proportion of patients who improve is higher for the drug group than for the placebo group.

Question 1 of 5
In a two-proportion z-test, why do we use a pooled proportion \( \hat{p}_c \) instead of the individual sample proportions?
Question 2 of 5
In a study, \( \hat{p}_1 = 0.45 \), \( n_1 = 100 \), \( \hat{p}_2 = 0.35 \), \( n_2 = 120 \). What is \( \hat{p}_c \)?
Question 3 of 5
A test yields \( z = -1.85 \) for \( H_a \colon p_1 - p_2 < 0 \). What is the p-value?
Question 4 of 5
Which condition check uses the pooled proportion \( \hat{p}_c \) in a two-proportion z-test?
Question 5 of 5
A two-sided test at \( \alpha = 0.05 \) gives a p-value of 0.038. Which conclusion is most appropriate?