STATS 499 Lecture Notes - Lecture 11: Relative Risk, De Havilland Mosquito, Reggiane Re.2000
Document Summary
We wish to study whether flight safety is related to plane size for commercial flights: measurements, measuring the size of the airplane. Could use data on mechanical malfunctions over number of miles or flights: this however, brings up the risk of aggregating data. Pilots have to file reports with anything that may be potentially unsafe during their flight: could access this data to have more data to work with. Simple random sample: this would not work because the number of unsafe flights is very small and if we considered 200 random flights, it is likely that they would all be safe. Look at large and small airplanes proportionately and then get safe and unsafe flights from both: there is no optimal ratio because it depends on some unknown factors, often used 1:2 or 1:3 ratio, 1:3 ratio. 30: let the probability of unsafe flights given large planes = p1 and the probability of unsafe flights given small planes = p2.