Consider testing the null hypothesis H0: b3 = 0 in a multiple regression with k = 5 explanatory variables.
(a) How could you do this with a partial F-statistic?
(b) Because you can also test H0 with the usual t- statistic, what must the connection be between a t-statistic and partial F-statistic for testing one slope?
(c) Pick a multiple regression. Compare the t-statistic for a coefficient with the partial F-statistic for removing that one variable. How are they related?