Читать книгу Germany's Freefall - Hermann Dr. Rochholz - Страница 7

“Don't believe any statistics you haven't faked yourself” – is the consensus of many people who don’t have a high opinion of statistics. This isn’t surprising since statistics is a branch of mathematics and not very much appreciated. You are more likely to reject what you do not understand. This is human nature. At the same time, you search for arguments in favor of this attitude in order to confirm your own personal prejudices. That, too, is human, but not effective.

However, this doesn’t make any sense because statistics create facts that can be used as evidence. On the contrary, in today’s society you are surrounded by statistics for any items because these are made of materials backed by statistics used to determine their material properties.

The validity of DNA expert opinions is pure statistics. Although a match between two DNA samples can thus never be 100%, it is on the order of 99.999995%, depending on the case; or a probability of 1 in 20 million. When you reject statistical evidence and wish to “argue” polemically, as at the beginning of this chapter, you then become inevitably in favor of abolishing DNA evidence and releasing sex offenders in particular. Does anyone really want that?

All scientific studies must be evaluated statistically. This is necessary to demonstrate to what extent the results of the study yield any valid results. When a study reveals a statistically significant result, this means that the result is not random but can instead be taken as “evidence”.

In the introduction to one of his books, the biostatistician, Bruce Weir, shows that the statistics of the discoverer of the basic principles of heredity, Gregor Mendel, are too good: The latter had let some “inappropriate” results fall by the wayside. It's indeed possible to validate statistics with statistics.