liveBook · Manning

Critical value

As explained in Timeless Algorithms, The Foundational Ideas

A critical value is a boundary value of a test statistic that separates the rejection region from the non-rejection region.

A critical value is the boundary of a test statistic’s rejection region. Values beyond that boundary are treated as sufficiently unusual under the null hypothesis to reject it. The boundary is set by the significance level before examining the test result. section 4.2.2 section 4.2.1

The decision rule depends on how the significance level is placed across the test’s tail or tails:

  • In a one-tailed test with α = 0.10, the upper-tail critical value is approximately z = 1.28; a z-score above it falls in the rejection region. section 4.2.2
  • In a two-tailed test with α = 0.05, 2.5% lies in each tail, and values beyond either critical value fall in the rejection region. section 4.2.1

Safety monitoring

Engineers compare an aircraft’s vibration z-score of 3.1 with the critical value 1.28. Since the statistic exceeds the boundary, they reject the null hypothesis of normal operation and take the aircraft out of service for inspection, accepting the risk of a Type I error. section 4.2.2

Beverage fill volumes

A sample t-statistic of approximately −2.24 falls beyond the lower critical value, approximately −2.09, for a two-tailed test with 19 degrees of freedom and α = 0.05. The manufacturer rejects the null hypothesis that the process meets its target average. section 4.2.1

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Explains how significance levels set one- or two-tailed boundaries and applies them to decisions in worked examples.
Examples
Uses aircraft vibration z-scores and beverage-fill t-statistics to illustrate one- and two-tailed decisions.
What it assumes
Requires a test statistic and significance level to establish the decision boundary before examining the result.
What you can do after
Apply a prespecified critical-value rule to decide whether a test statistic falls in the rejection region.
How it connects
Connects critical values to hypothesis testing, null hypotheses, rejection regions, significance levels, p-values, and Type I errors.

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.