Clandestino wrote...
Soylentgreen wrote...
Given an identical situation, what % of professional pilots (or perhaps '3 man groups of pilots') would flub it and crash the plane?
Those who have been paying attention know that between Nov 12 2003 and Aug 07 2009, there were 37 recorded cases of unreliable airspeed on A330/340 worldwide. 36 of them ended without damage to aircraft or injury to anyone. One ended up in airframe write-off and death of all on board.
Study you proposed has already been done. Results are in the final report. Your notion that:
A big part of this was the human-machine interface, which did an extremely poor job of letting the pilots know what was actually going on.
...is not confirmed.
Thanks to those who liked my post
You are welcome, even as I liked it as a very good example of bad science.
Let's assume that the 1 out of 37 failure data that you quote was a true experiment (it wasn't, it's an observational quasi-experimental design, but let's ignore that for now).
1 out of 37 is about 2.7%, meaning that based on this data, roughly 3% of crews in identical situations would crash the plane.
Of course, we know from basic statistics that small sample sizes are problematic. Plugging the #s into a confidence interval calculator gives us a margin of error of roughly +/- 7%, at the 99% confidence level.
In plain English, with a sample size of 37, we can only say with confidence that actual rate of crashing in this situation will be between 0% and 10%.
Care to explain your "Bad Science" comment? If you don't have doctoral-level statistics knowledge and don't understand this, I'd be happy to explain in more detail?
I think your point that "the study...has already been done..." is valid, only if you accept that results between 0% and 10% are acceptably precise. Do you?