Chapter 10: Hypothesis Testing with Two Samples
📄 Key Definitions:
Two-Sample Hypothesis Testing: A procedure to compare two population parameters
(means or proportions) using independent or paired samples.
Independent Samples: Two samples that are not related (e.g., men vs. women).
Dependent Samples (Matched Pairs): Samples are related or paired (e.g., before-and-
after measurements).
Null Hypothesis (H₀): Assumes no difference between the two population parameters.
Alternative Hypothesis (Hₐ): Suggests a difference exists.
✅ General Steps:
1. 1. State H₀ and Hₐ.
2. 2. Choose significance level α (e.g., 0.05).
3. 3. Determine test type (z-test or t-test; proportions or means).
4. 4. Calculate the test statistic.
5. 5. Find p-value or critical value.
6. 6. Compare with α.
7. 7. Make a decision and interpret the result.
📊 For Two Means (Independent Samples)
Case 1: σ² Known (z-test, large samples):
z = ((x̄₁ - x̄₂) - D₀) / √(σ₁²/n₁ + σ₂²/n₂)
Case 2: σ Unknown (t-test, small samples):
t = ((x̄₁ - x̄₂) - D₀) / √(s₁²/n₁ + s₂²/n₂), df ≈ min(n₁ - 1, n₂ - 1)
📊 For Two Proportions (Large Samples)
Pooled proportion: p̂ = (x₁ + x₂) / (n₁ + n₂)
Test statistic: z = (p̂₁ - p̂₂) / √[p̂ (1 - p̂ )(1/n₁ + 1/n₂)]
📅 Assumptions/Conditions:
Samples must be random and independent.
For t-tests: populations must be approximately normally distributed (or n ≥ 30).
For proportions:
- np ≥ 5 and nq ≥ 5 for both groups.
🔍 Matched Pairs (Dependent Samples):
Use differences: Let d = x₁ - x₂
Then apply one-sample t-test to d̄:
t = (d̄ - μ₀) / (s_d / √n)
🌍 Interpretation:
If p-value < α → reject H₀ → conclude there is a difference.
If p-value ≥ α → fail to reject H₀ → not enough evidence to show a difference.
