Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

How to convince someone of statistical probibilities

  1. Oct 28, 2005 #1


    User Avatar
    Gold Member

    Ok im having a really annoying time with my father here. Now statistically, you are much more likely to get in an accident in a car then in a small airplane per mile traveled. Now my father thinks that its better to say that its more dangerous to have a small airplane since you are more likely to get in an accident based on the number of times you start your engine.

    Now i say, the statistic should be based on miles, he says on startup and I'm almost positive my statistic is the correct basis since people's accidents occur while driving and not when they start-up their car. Extrapolate the idea out and someone who drives 100000 in one trip has a faaaaaaaaaaaaar larger chance of getting in an accident then someone who turns his car on once and drives across the street and back.

    Now i want to really verify that my statistical basis is correct... and if it is, how can you convince someone that you are right.

    This may be more reasonably put in the GD section.
  2. jcsd
  3. Oct 28, 2005 #2


    User Avatar

    the number of planes that will crash will depend on the number of times the engine has started, by the nature of their crashes (AFAIK, a longer trip does not imply a greater risk, since the plane would have a problem due to something else, ie mechanical failure when starting).
    For cars, however, it is the opposite; a longer trip brings a greater risk.

    If you compare them using only the statistics from the miles travelled, then you dont take into account the fact that the planes' accidents do not depend on the number of miles travelled, but rather on the number of flights.
  4. Oct 28, 2005 #3


    User Avatar
    Gold Member

    Yah but the fact it has a problem is the thing. When it has a problem, theres a possibility of a crash (thus, an accident).
  5. Oct 29, 2005 #4


    User Avatar
    Science Advisor
    Homework Helper

    Think takeoffs and landings! :)

    People who want to convince you that flying is safe use the "passenger-mile" model. Exposure time (time in transit) and number of trips (takeoffs and landings) are also valid measures but the airlines and FAA don't want to alarm people. I don't think there is a strict "right or wrong" on this issue.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook