Depends on how it is being loaded.
A simple beam supported at both ends and loaded with a single weight in the middle will result in compression of the fibers on the "top" and tension in the fibers on the "bottom". The stress (load per unit area) on the extreme fibers will always be the highest, basically due to leverage about the center of gravity of the piece in question, and the further apart these fibers are the higher the Moment of Inertia, or ability to resist bending for a given strength of material. Kind of like a lever; the further out on the lever there is less force required to resist a given torque load, and therefore the fibers either don't need to be as strong, or they can take a bigger load.
So if your concern is that you want to be able to stand on the roof of your foamie, then you probably would want a material that is good in tension on the ceiling, and one that is good in compression on the roof. For the same materials, if you want them to handle bigger bending loads in the long axis, place the fibers further apart (thicker). If the walls supporting your roof are further apart (6, 7, 8 ft wide); thicker.
If, however, you are building a tanker to haul 500 pounds of beer in bulk form, then the pressure is coming from the inside and you want the tensile member on the outside and the compressive member on the inside. (Or maybe you plan on leaning up against a wall as a back rest, or standing on your floor.)
Bunk beds or a hammock? Tension on the inside of the wall from the weight of the person sleeping trying to pull the wall in, compression on the outside.
Of course this is all just static theory.
The reality of a simple foamie construction is that the sock not only provides an exterior skin, it also is a means of fastening the main panels together at the seams. And the geometry of this joint makes it far easier to apply effectively from the outside (it is easier to apply tension pulling the joints together by stretching the fabric over an outside corner, than it is to cram it into an inside corner, which by the nature of the application method would be tending to separate the joint. When looking at the big picture, at some point the cocoon, or overall shell construction becomes the other side of the stressed member (assuming that there are at least some shear walls (bulkheads, structural cabinet frames, roof spars, front and rear walls, etc.) spanning between the side walls.
My thought is that fiberglass weave and epoxy become more monolithic when cured (better in both tension and compression) than cotton fabric and glue or Latex, so might work equally well on the inside. Whereas the glue and cloth on the inside is doing more to protect the soft foam from local damage due to incidental contact from the occupants. Again, my perspective here is for a small simple teardrop structure. Larger panels will need stronger tensile and compressive qualities for both the interior and exterior skins (or the panels must be thicker) because the spans are greater and the loading can be expected to be more diverse, coming form either direction, or both.
But again the reality is, who wants to be stuck inside a little box applying epoxy and glass (ick). That and the warm appearance of wood is the reason I plan to use thin plywood on the inside and canvas/glue/paint on the outside (but I am not building a big camper, nor an ultra light).
Just my ramblings on the topic, take them for what they are worth.