The P-38 has 15 degrees of dihedral, according to what Warren (Bodie) told me. A little more dihedral (actually, I think 5 degrees is more like it, ten is too much) would have raised the booms above the cockpit nacelle and center wings a little more, which would have put the elevators in cleaner air. That does NOT change compression much, except for how air would travel over the wings with more dihedral. It would possibly have slowed the roll rate a little more. The wingspan is what made the P-38 so slow to roll, and it was so stable, if a little more dihedral made it less stable it would not have hurt it.
The larger radius would have helped a little more. But angling the leading edge of the center wings so that the edges would have angled back from the center nacelle to the booms would have caused the air to flow towards the booms and slowed the air down, which would have helped mitigate or lessen the effects of compression and possibly raised the critical Mach number by a fair margin.
Compression is not the same as elevator flutter, which is what the USAAF thought the problem was. Compression does not lock the elevator, what it does is change the way the wing generates lift, and moves the center of lift towards the rear. It was also call "Mach Tuck", because it caused the nose to "tuck under".