Two-Tailed Tests: Testing for Any Difference
You'll learn how two-tailed tests detect differences in either direction by splitting your significance level between both extremes.
What Is a Two-Tailed Test?
A two-tailed test is used when you care about detecting any difference from the null hypothesis—whether the true parameter is larger or smaller than your hypothesized value. Unlike one-sided tests (which look for change in only one direction), two-tailed tests keep watch on both extremes.
How Alpha Gets Split
Remember that your significance level (α) represents your tolerance for Type I error—rejecting a true null hypothesis. In a two-tailed test, you divide this alpha equally between both tails of the sampling distribution:
- Left tail: α/2 (e.g., 0.025 if α = 0.05)
- Right tail: α/2 (e.g., 0.025 if α = 0.05)
This means evidence in either tail can lead to rejection. If your test statistic falls into the extreme 2.5% on the left OR the extreme 2.5% on the right, you reject the null hypothesis.
When to Use Two-Tailed Tests
Use a two-tailed test when your alternative hypothesis is non-directional. For example:
- H₀: μ = 100
- H₁: μ ≠ 100 (could be larger or smaller)
You're not predicting which way the difference goes—you just want to know if there's any difference.
The P-Value Connection
For two-tailed tests, the p-value represents the probability of getting a test statistic as extreme as yours in either direction. If your test statistic is in one tail, you typically double the one-tail probability to account for both possibilities.
Key Takeaway: Two-tailed tests split your alpha (α) equally between both tails of the distribution, allowing you to detect deviations in either direction while maintaining your overall Type I error rate.