Tuesday, September 18, 2007

After you plot your data...

It still amazes me that I'm getting an A in my statistics class. I flip back through pages of notes and see all of these little diagrams with shaded bits and formulas with numbers and odd notations and since it's in my writing - I guess I must have done all that. I am giving a major portion of the credit on this to the class itself - it's really well written and breaks down all the information into small, manageable sections. It's really designed to make absolutely sure you fully understand each concept before you move on to the next one, with non-credit self-checking quizzes and a tutor available by an 800 number or email. But wow - am I ever grateful that this class turned out to be one of the best of the bunch. (Unlike certain OTHER classes which we will discuss at a later date - perhaps when the appropriate beverages are within reach.) I'm a little nervous about the exams, because while it's one thing to select the right formula, plug in the numbers, and then solve it correctly - there's supposed to be a section of problems that require you to interpret and discuss the results. This I find difficult sometimes! But maybe that's because I try to read too much into it, and maybe sometimes, the answer is very simple.

During World War II a statistician named Abraham Wald worked on war problems and some of the statistical methods he invented were militiary secrets until the war ended. But some of the methods he used to address problems were not that complicated. When the militairy asked him where extra armour plating should be added to airplanes, Wald studied the location of enemy bullet holes in planes returning from combat. He plotted the locations on an outline of a plane. As the data accumulated, most of the outline filled up. Wald then told the militairy to put the armour in the few spots where no bullet holes were marked. That's where bullets hit the planes that didn't make it back.

No comments: