Not trying to argue w/ anyone, but this is a pretty good topic to philosophize.
Nice Q & A info, I would agree that the self-tapping thread could have a tendency to back out more, when compared to the machine screws.
However, it just seems that the machine screws, when inserted into the nylon, would just rip out. True, they have a tighter thread pattern than the self-tapping. This is why the machine screw will not back out.
But, what about the material they are tapping into? It is relatively soft, like wood, definitely not like metal. So in theory, you have more threads (which does help the back out problem), but you are also reducing the grip factor. Thus, you will probably strip the nylon with the machine screws. Look at the threads for screws that go into other soft material such as wood. They definitely don't use machine screws for putting together houses, decks, piers, speaker boxes, etc....
Look at the spaces that are b/t the threads. What do you think would be stronger, a piece of nylon that is this wide /...../ (self-tapping) or a piece of nylon that is only this wide /../ (machine) ?
The thin pieces of nylon that would be b/t the threads of the machine screw would be weaker, in comparison to the thicker pieces of nylon from the sef-tapping screw. By using a machine screw in soft material, you are basically putting a plug, or a nail in it. True, it solves the back out issue, but it seems that it would strip faster and also provide less support.
Another thing to consider is that the screws are going into the diff housing and it is a high stress area. So imagine this example. The thickness diagrams (above) represented a similar high stress area such as a dam holding back water. I would rather live below a dam that was this thick /...../ than this thick /../