We seem to get this kind of question pretty regularly so let me try to address it general terms and if you still need help seeing what's going on in your file send it over (use the Secure File Upload link below).
The forces are indeed being redistributed by the diaphragm to other frames as you suggested. In my opinion it's perfectly reasonable and accurate assuming the diaphragm is rigid. We sometimes suggest that users model the diaphragm as semi-rigid if they are uncomfortable with the magnitude of the force transfer, but even with semi-rigid diaphragms frame shear transfer will occur. The only way to eliminate it completely is using flexible (or pseudo-flexible) diaphragms in Ram Frame.
Why does this happen? The following wiki does a nice job of explaining why the shear forces in two identical frames (or walls) may not be the way you expect when one frame is taller than another and they are linked by a rigid diaphragm:
Image may be NSFW.
Clik here to view.
The same principles apply in comparing two frames attached to the same diaphragms where one frame is wider or stiffer than the other. In these cases the % of the total shear resisted by each frame is not linear from story to story, even if the stories are all identical. Here's an image from a study I did long ago to put this in perspective:
Image may be NSFW.
Clik here to view.
Another way to put it is this way; diaphragms ensure that the frames have equivalent lateral displacement, but the diaphragms do not limit vertical displacement of the nodes, thus diaphragms can transfer shear forces from one isolated frame to another, but they cannot transfer moments or vertical forces between those frames (If the frames are liked by frame beams with fixity or two-way decks with out-of-plane stiffness, that's a whole different situation).
