liveBook · Manning

Critical region

As explained in Timeless Algorithms, The Foundational Ideas

A rejection region is the set of test-statistic outcomes for which a hypothesis-testing procedure rejects the null hypothesis.

A critical region, or rejection region, is the set of test-statistic values that lead to rejection of the null hypothesis. It turns a statistical test into a decision rule: if the observed statistic falls in the region, reject the null; otherwise, fail to reject it. section 4.1.5

The significance level α specifies the maximum acceptable probability of a Type I error: rejecting the null hypothesis when it is true. The critical region translates that probability into values of the statistic using its sampling distribution under the null. In a two-tailed test at α = 0.05, the extreme 5% is divided equally between the two tails; a one-tailed test places the region on one side. section 4.1.5

Figure 4.1 Illustration of critical regions in a two-tailed hypothesis test at the 5% significance level. The curve represents the distribution of the test statistic under the null hypothesis. In a two-tailed test, the rejection regions are placed symmetrically in both extremes of the distribution, with 2.5% in each tail, so that results unusually far from the center in either direction would lead to rejecting the null. By contrast, a one-tailed test would place the entire 5% rejection region on only one side of the distribution, used when deviations are meaningful only in a single direction. This figure provides one example of how critical regions formalize the decision rule in hypothesis testing.

The figure shows a two-tailed critical region with 2.5% in each tail and contrasts it with a one-tailed rule. section 4.1.5

Critical values mark the boundaries of the region. The same decision can be expressed using a p-value: the probability, assuming the null hypothesis, of a statistic at least as extreme as the observed one. Reject when p ≤ α. Set the rule in advance so that the observed data do not determine the standard of evidence. section 4.1.5

Safety monitoring

For unusually high aircraft vibration, a one-tailed z-test uses α = 0.10 and an upper critical value of approximately 1.28. An observed z-score of 3.1 lies in the rejection region; its one-tailed p-value is approximately 0.001. The null hypothesis of normal operation is rejected, and the aircraft is removed from service for further inspection. The decision still carries a risk of a Type I error. section 4.2.2

A discrete threshold

In a trial of 20 patients with a null recovery probability of 50%, a one-tailed binomial test at α = 0.05 rejects for 15 or more recoveries. Fourteen recoveries fall short of that region, with p approximately 0.057. Failing to reject does not establish that the drug is ineffective or that the null hypothesis is true; it means the evidence did not meet the predetermined threshold. section 4.2.3

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Defines rejection boundaries and works through one- and two-tailed tests, including a discrete binomial threshold.
Examples
Uses aircraft vibration monitoring and a small clinical trial to illustrate rejection and failure to reject.
What you can do after
Evaluate whether an observed statistic warrants rejection using a predetermined critical region or equivalent p-value comparison.
How it connects
Connects significance level, Type I error, sampling distribution, critical values, and p-values.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Timeless Algorithms, The Foundational Ideas. The section links point to the parts of the book it draws on.