How could random variation affect that statistic?
If the null hypothesis is true, then z or t is near zero (chi-square and F are small). If the assumptions are met, the test statistic has a specific distribution-that is, we can assess how “typical” a value is. Extreme values of the test statistic reflect larger differences (stronger relationships) and thus favor the alternative hypothesis. All this comes down to: determine how you’ll calculate a p-value.
