And to think I had a successful research career before taking up aviation. Of course, you didn't know that, but disrespectful potshots are easy, eh?
Same here but my line was mechanical engineering.
I don't have a clue about the US safety databases but the UK ones carry the airframe manufacture date for any accident.
You could just go for the types which have been certified in the last 15 years and then factor for the fleet size.
As I said before the way you factor the data makes a huge difference.
Take off and landing is by far the most risky stages of flight so by removing the instructional flights and commercial you have removed the most exposure to risk flights. ie the ones that do the most take off's and landings in instructional and the commercial are the ones which are exposed to the barely legal wx. So to me anyway its is wrong to discount that data.
Another way you could look at the data is to look at the number of flights irrespective of the length of the flight. But again that won't produce the effect your looking for as C150/C152's and other training flights are banging out the sectors a day without having many problems.
And it would change my view. Just like as BPF says I was a teach them spinning type of instructor and when looking at the data am now firmly in the teach them how not to get in the situation of a likely spin.
Although for advanced handling for experienced pilots post PPL when done by an aero's instructor that knows what they are doing and in a machine that's fit for purpose I think its good training even if only done once.
And GA to me is anything with a single pilot, one engine and under 5700kg.