Mathematics / Mathematik / Matemática
Posts 1-3 of 3
-
Ansgar Seyfferth Group moderatorThe company name is only visible to registered members.Selection bias
Make sure your observations are representative when using them for statistical inferences!
http://fb.me/trIIqVRG
- 17 Oct 2010, 8:05 pm
-
Post visible to registered members
-
Ansgar Seyfferth Group moderatorThe company name is only visible to registered members.
Re^2: Selection bias
Sure, you are right.
What I meant is that a large majority of people would intuitively consider the areas with most observed bullet impacts as the most exposed ones of the aircraft and where hence more armor should be placed, as if the observed (i.e. returned) aircrafts were a representative selection, without taking into account that it excludes precisely the most relevant planes, i.e. the ones which suffered the fatal bullet impacts and as a consequence did not return.
By the way, here you have a publication from Marc Mangel and Francisco J. Samaniego about Abraham Wald's work on aircraft survivability:http://people.ucsc.edu/~msmangel/Wald.pdf
- 18 Oct 2010, 12:14 am
