In statistics, can you use nested F-test in deciding an ARIMA model

Let's say we have 2 models, MA(2) and MA(3). Both models are significant overall. In each model, all of the terms are also significant individually. Can we use nested F-test to determine whether or not it is worth including the MA(3) term into the model? Thank you.

QUICK 10 POINTS!!

I don't believe that you can. I think that that only applys to models with independance