My definition of Fatality Rate is the number of accidents with at least one fatality, vs. the total number of accidents.
Certainly, but unfortunately, that's difficult data to extract. Plus, of course, accidents are random events...a fatality might depend on whether the passenger is leaning over to tie his shoe, etc.
I certainly understand I can't *prove* the theory that wing position affects survivability. I'm attempting to find some correlation by using thousands of aircraft in each configuration. The homebuilt accident record provides about 3800 accidents over the 19-year period my database runs.
Obviously, the speed at impact is a very big driver...energy increases by the square of the impact velocity. I'm hoping to use cruise speed to differentiate...but, of course, few airplanes actually CRASH at cruise speed. I'm using cruise speed just to indicate the approximate operating range of the aircraft. An RV-6 approaches faster than the cruise speed of Frank's Nieuport, but by using cruise speed as the factor the difference in capability is acknowledged.
One drawback is that there ARE no quantities of high-performance high-wing homebuilts. There are a few Aerocomps, even fewer Stallions. So it's hard to get data in the higher-performance regime.
There is some correlation in the production-aircraft world. The Cessna 172 has a fatality rate of 12.%, while the PA-28 (through the -181) it's 19%. The Bonanza and the Cirrus both have rates in the low '30s, while the Cessna 210 is 20%. However, the 210 has a lot of landing-gear-related and fuel starvation accidents, which usually aren't fatal.
Ron Wanttaja