Hypothesis testing

Also known as: Statistical test, Significance test

Hypothesis testing decides whether data are compatible with a claim about $\theta$. Specify null $H_0$ and alternative $H_1$, choose a test statistic, and reject $H_0$ when the statistic falls in a critical region (or when the $p$-value $<\alpha$).

Type I error: reject a true $H_0$ (rate $\alpha$). Type II: fail to reject a false $H_0$. Power $=1-$ Type II. Regression slope tests are usually two-sided $H_0:\beta=0$ via a $t$-test.

← Back to CS Glossary