Hypothesis testing: t-tests

t-tests are used in hypothesis testing to compare two population means.

The null hypothesis is Ho: μ₁ = μ₂ and the alternative hypothesis is Ha: μ₁ ≠ μ₂.

Different procedures are used to calculate the t-value for related and independent samples. In this applet the Data panel has charts describing the samples taken from two populations, and the t-test panel displays a t distribution with the t-value, p-value and critical value. Note that the critical value is computed from the significance level and is used to either accept or reject the null hypothesis.