Mathwords logoMathwords

Sampling Distribution of p̂₁ − p̂₂ — Definition, Formula & Example

The sampling distribution for the difference between two sample proportions shows every value p^1−p^2\hat{p}_1 - \hat{p}_2 could take, and how likely each one is, when you repeatedly draw a random sample from each of two populations. It tells you how much the gap between two sample percentages naturally varies by chance alone.

For independent random samples of sizes n1n_1 and n2n_2 from populations with true proportions p1p_1 and p2p_2, the statistic p^1−p^2\hat{p}_1 - \hat{p}_2 has mean μp^1−p^2=p1−p2\mu_{\hat{p}_1 - \hat{p}_2} = p_1 - p_2. Its standard deviation is σp^1−p^2=p1(1−p1)n1+p2(1−p2)n2\sigma_{\hat{p}_1 - \hat{p}_2} = \sqrt{\frac{p_1(1-p_1)}{n_1} + \frac{p_2(1-p_2)}{n_2}}, valid when each sample is at most 10% of its population. If n1p1n_1p_1, n1(1−p1)n_1(1-p_1), n2p2n_2p_2, and n2(1−p2)n_2(1-p_2) are all at least 10, the distribution is approximately Normal.

Key Formula

μp^1−p^2=p1−p2,σp^1−p^2=p1(1−p1)n1+p2(1−p2)n2\mu_{\hat{p}_1 - \hat{p}_2} = p_1 - p_2, \qquad \sigma_{\hat{p}_1 - \hat{p}_2} = \sqrt{\frac{p_1(1-p_1)}{n_1} + \frac{p_2(1-p_2)}{n_2}}
Where:
  • p1,p2p_1, p_2 = true population proportions for populations 1 and 2
  • n1,n2n_1, n_2 = sample sizes drawn from populations 1 and 2
  • p^1,p^2\hat{p}_1, \hat{p}_2 = sample proportions computed from each sample
  • μp^1−p^2\mu_{\hat{p}_1 - \hat{p}_2} = mean (center) of the sampling distribution
  • σp^1−p^2\sigma_{\hat{p}_1 - \hat{p}_2} = standard deviation (spread) of the sampling distribution

Worked Example

Problem: At School A, 60% of students own a bike. At School B, 50% do. Each school has more than 1,000 students. You take independent random samples of 100 students from each school. Describe the sampling distribution of p^A−p^B\hat{p}_A - \hat{p}_B. Then find the probability that the School A sample has a smaller proportion of bike owners than the School B sample.
Step 1: Check the conditions. Each sample of 100 is less than 10% of its school. The expected counts are 100(0.6)=60100(0.6)=60, 100(0.4)=40100(0.4)=40, 100(0.5)=50100(0.5)=50, and 100(0.5)=50100(0.5)=50. All are at least 10, so the shape is approximately Normal. The mean is the difference of the true proportions.
μp^A−p^B=0.6−0.5=0.1\mu_{\hat{p}_A - \hat{p}_B} = 0.6 - 0.5 = 0.1
Step 2: Find the standard deviation. Add the two variances first, then take the square root.
σ=0.6(0.4)100+0.5(0.5)100=0.0024+0.0025=0.0049=0.07\sigma = \sqrt{\frac{0.6(0.4)}{100} + \frac{0.5(0.5)}{100}} = \sqrt{0.0024 + 0.0025} = \sqrt{0.0049} = 0.07
Step 3: School A's sample proportion is smaller exactly when p^A−p^B<0\hat{p}_A - \hat{p}_B < 0. Standardize 0 and use the Normal model.
z=0−0.10.07≈−1.43,P(Z<−1.43)≈0.076z = \frac{0 - 0.1}{0.07} \approx -1.43, \qquad P(Z < -1.43) \approx 0.076
Answer: The sampling distribution of p^A−p^B\hat{p}_A - \hat{p}_B is approximately Normal with mean 0.1 and standard deviation 0.07. The probability that School A's sample proportion is smaller than School B's is about 0.076, or roughly 8%.

Why It Matters

This distribution underlies the two-proportion z-interval and the two-proportion z-test in AP Statistics. Medical researchers use it to decide whether a drug's success rate truly beats a placebo's. Pollsters use it to judge whether a gap between two groups' approval ratings is real or just sampling noise.

Common Mistakes

Mistake: Adding the two standard deviations, or subtracting the variances because the statistic is a difference.
Correction: For independent samples, variances always add, even for a difference. Add p1(1−p1)n1\frac{p_1(1-p_1)}{n_1} and p2(1−p2)n2\frac{p_2(1-p_2)}{n_2}, then take the square root. In the example, adding the separate SDs (0.049+0.050.049 + 0.05) gives about 0.099 instead of the correct 0.07.

Related Terms