You have a couple choices for the 16:9 footage.
You could center-punch it. That is, show only a 4:3 window view of the center of the 16:9 picture. This would be like when you rent a movie that has been made “full-screen” by the pan-and-scan process. Easy to do in any NLE with just a wipe or crop.
You take a risk with a center-punch not containing all of the vital picture information due to he way the original was framed. Which is one reason why I hate pan-and-scan movie rentals.
OTOH, you have an advantage that you can shift that underlying large wide picture around on it’s own layer to try and get the best 3:4 window out of it, as well as to add pans and tilts and even zooms that were not in the original. This method also doesn’t hurt the final picture quality as much as scaling might.
When the original shooter is composing his or her frame, some like to shoot to “protect” the 4:3 ratio. That is to say, they shoot wide but arrange the key parts of the shot to still fall into a 3:4 center spot. You do that mostly when you are not sure but suspect that this will get a lot of use in 4:3. It is definitely an aesthetic choice you must make, and some folks say you are crippling the ability to make a better picture in wide screen if everything must be shot for “protection”. A policy discussion on this is best held and agreed to before a single frame is shot.
Scaling is the other choice, re-sizing the entire frame. I think generally this is not the way to go, because it needs to radically shrink your widescreen picture to fit into a letterbox or pillarboxed window, which only keeps drawing attention to the difference in proportion. At the same time you make everything in that shot too tiny by comparison, and add distracting black bars.
If the majority of the footage is wide screen, I say, conform the 4:3 footage to match the widescreen. There are some tricks that let you get away with that. If most of the footage is 4:3 ratio, I would rather see the wide screen cropped to mach the smaller framing.