Invalid Statistics

The way drugs are tested in the USA makes no sense statistically. The drug company is permitted to hide any trials that did not produce promising results. You are not permitted to cherry pick results if you want statistically valid results. Assume Nestlé want to market its Smarties candies as a miracle sports enhancer. They could find a fat kid and ask him to shoot a basketball foul shot after popping a blue Smarty, over and over and over. They film all this and cut out all the film where he failed to make the shot. The cherry-picked result makes the kid on Smarties look like Magic Johnson. That is what drug companies are getting away with when they are permitted to hide unsuccessful trials.

~ Roedy (1948-02-04 age:70)