I see your thinking. You approximate the bunk as two edges over which the plate is bent. In the ideal case, the hull over the middle of the bunk wouldn't even touch. Then each "edge" acts independently and their separation doesn't change the picture much.
On a DS1, the seats provide a support about 1' or so from the centerline. If your bunks are wide enough to reach that point then the hull is no longer an ideal plate.
The other way you can get away from that ideal model (even for a DS2) is if you assume that the bunks are padded and that the hull plate is not totally flat. Both may allow a more gradual transfer of load and that may well be different for wider bunks.
Finally, bunks should be made both longer and wider. That puts more of the support closer to the edge of the hull panel, which might mean it's closer to a support (e.g. transom) and acts less like an ideal plate.
Also, if your bunk is more like a square than like a beam (and the end of a wide bunk would look like a square) that itself would increase the area of the plate over which the load is distributed and change the load from linear to something more 2D (which may also have an effect).
I'd be interested to see what conclusion you would draw from the theory of plates, so I'm hoping you can find your copy of Timoshenko.