We compare two sample variances to determine whether the difference between them is probably not due to random sampling error when we perform an F-test, or a variance ratio test. The null hypothesis for the F-test is that the population variances are equal, while the alternative hypothesis is that they are not equal.
Hope this helped!