Course contentsShow
← Data Science
Lesson 325 of 2,1459. Hypothesis Testing FoundationsFree lesson

The Rejection Region

Defining the critical region in the test statistic's distribution where we reject the null hypothesis.

The Rejection Region

What You'll Learn

You'll learn how to define the rejection region — the specific zone in your test statistic's distribution where the evidence is strong enough to reject the null hypothesis.

The Core Idea

Think of the rejection region like a "no-go zone" in your test statistic's distribution. When you conduct a hypothesis test, you calculate a test statistic from your sample data. This statistic has its own sampling distribution under the assumption that the null hypothesis is true.

The rejection region (also called the critical region) is the set of values that are so extreme — so unlikely under H₀ — that if your test statistic lands there, you reject the null hypothesis.

How It Works

The rejection region is directly tied to your significance level (α). For example, with α = 0.05:

  • Two-sided test: The rejection region sits in both tails of the distribution, with 2.5% in each tail
  • One-sided test: The entire 5% sits in one tail (left or right, depending on your alternative hypothesis)

The boundaries of these regions are marked by critical values. If your calculated test statistic falls beyond these critical values (into the rejection region), you reject H₀.

A Simple Analogy

Imagine a dartboard where the bullseye and nearby rings represent "typical" results under the null hypothesis. The outer edges represent extreme, unlikely outcomes. The rejection region is like drawing a line and saying, "If my dart lands outside this line, the result is too extreme to believe the null hypothesis is true."

Key Takeaway: The rejection region defines the extreme values of your test statistic where you have sufficient evidence (at your chosen α level) to reject the null hypothesis; landing in this region means your sample result is too unusual to attribute to random chance alone.