Used to determine whether there are significant differences between observed and expected frequencies in categorical data.