But - in what was probably the most controversial finding - the probability of an accident as the result of a non-common cause engine failure was greater on a quad or a tri than a twin.
I suspect that you haven't expressed that the way you intended.
A "a non-common cause engine failure" implies more than one engine failing, right ?
So do you mean the relative probability of multiple engine failures on a quad vs a twin ?
Or do you mean,
given multiple engine failures, the relative probability of an accident ? Nobody could possibly argue that that's higher with 2 remaining engines than with none.
Looks like we're back in compound/conditional probability territory - I'd really like to see some numbers, if anybody has some.