The binary variable interaction regression

(A) can only be applied when there are two binary variables, but not three or more.
(B) is the same as testing for differences in means.
(C) cannot be used with logarithmic regression functions because log(0) is not defined.
(D) allows the effect of changing one of the binary independent variables to depend on the value of the other binary variable.