@Setanta,
Setanta wrote:
Do you have equivalent data on the number of flights per annum? It only becomes significant, really, if one can compare the number of fatalities and injuries to the number of people using air transport. (Which assumes that air transport safety is the issue.)
I understand the thought that comparing the number of world-wide flights to crashes can give one a rough idea of the odds of crashing. However, would a more accurate correlation be obtained by comparing the number of crashes to total miles flown world-wide, since their might be a correlation to longer flights and greater odds of a problem? For that matter, if one is concerned about one's safety on a particular trip, would it not be the best statistic to compare number of crashes for that type of trip only (e.g., number of crashes for flying domestically, or number of crashes flying overseas)? And then broken down by season, and within season by day flights, and night flights. Getting into respective weather conditions may not be feasible, since weather affects a flight during take-off, landing, and the flight itself.
And then there is the non-obvious, like whether maintenace was out-sourced, or done by airline personnel.
My point is that there are many variables, and to put one's safety in the simplest of statistics (total flights, versus total crashes, per year) is not the best use of statistical comparisons, I believe. Or, at least a histogram with classes for each criterion, broken down to total flights, versus total crashes for each criterion). Then one can choose a flight with the safest criteria (possibly: summer, day flight, domestic, fair weather, company personnel doing maintenance).
Oh, then there are the variables for the respective take-off and landing airports. That could be a study in itself, perhaps.
Choosing to fly safely is no different, I believe, than handicapping a horse race. The problem is a horse race gives the bettor much information to do an analysis. Not so with flying.