**A test for statistical significance that relies on a comparison of the** ratio of two mean square errors. For example, one can use the ratio of "mean square due to the regression" to "mean square due to error" to test the overall statistical significance of a regression model. *F = t*^{2} (see t-test).