Course contentsShow
← Mathematics
Lesson 3037 of 3,12159. Statistics — Descriptive and InferentialFree lesson

Two-Sample t-Test: Independent Samples with Equal Variances

Test whether two independent populations have equal means assuming equal variances (pooled t-test).

Two-Sample t-Test: Independent Samples with Equal Variances

What you'll learn: How to test whether two independent groups have the same population mean when their variances are equal but unknown.

The Core Idea

Imagine comparing average test scores from two different classrooms, or measuring the effect of two different fertilizers on plant growth. When you have two independent samples and want to know if their population means differ, but don't know the population variances, you use a pooled two-sample t-test.

This test assumes:

  1. Both samples are independent (one doesn't affect the other)
  2. Both populations have equal (but unknown) variance
  3. Both populations are approximately normal, or samples are large enough

How It Works

Since we assume equal variances, we combine (pool) information from both samples to estimate the common variance more accurately. The pooled variance is:

sp2=(n1−1)s12+(n2−1)s22n1+n2−2s_p^2 = \frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}

where n1,n2n_1, n_2 are sample sizes and s12,s22s_1^2, s_2^2 are sample variances.

The test statistic is:

t=(xˉ1−xˉ2)−(μ1−μ2)0sp1n1+1n2t = \frac{(\bar{x}_1 - \bar{x}_2) - (\mu_1 - \mu_2)_0}{s_p\sqrt{\frac{1}{n_1} + \frac{1}{n_2}}}

Under the null hypothesis H0:μ1=μ2H_0: \mu_1 = \mu_2, this follows a t-distribution with df=n1+n2−2df = n_1 + n_2 - 2 degrees of freedom.

When to Use It

Use this test when comparing two independent groups (like treatment vs. control, or two different brands) where you believe the variability is similar in both populations. If variances appear very different, you'd use Welch's t-test instead.

Key Takeaway: The pooled two-sample t-test combines variance information from both independent samples to test whether their population means are equal, giving more power when the equal-variance assumption holds.