Tuesday, May 1, 2012

Baby names for statisticians

In my opinion most scientists, or - at least - probably most women scientists do not really like statistics. Or maybe, without generalization I can say that probably 85% of scientists use statistics (and think it is useful) but hate to do it. Or am I the only one who is often in trouble with statistical data treatment? Whenever it comes to statistics I have my own joke about it (and about its usefulness) as follows.
- How probable is it that I am Napoleon?
- ???
- 50 %.
- ??? (She must have gone mad.)
- No, it is 50 %, because either I am Napoleon or I am not.
Such problems with statistics are for instance well known for all pregnant women who make some genetic tests that provide some probability data... What is the significance/usefulness of a statistical test/method that is only 87% correct, that gives you 2-3% false positive results, and gives you a final estimation saying that your baby has some health or developmental problems with 1:23 probability. Science helped you a lot again...

1 comment: