0001以下、5ちゃんねるからVIPがお送りします
2022/02/27(日) 22:39:19.717ID:MN3kY0Xh0@agreement Ahypothesis Bindication Cparadox
There are seemingly correct logical mistakes in the world, which are sometimes a problem in the field of statistics.
Consider comparing the therapeutic effects of two drugs, A and B.
For example, A was found to cure a disease with a probability of 9 in 10 males and 30 in 100 females. On the other hand, B was cured with 60/100 males and 1/10 females.
In both cases, it can be said that A is superior to B when compared to men and women respectively.
However, when considering a mixture of men and women, the therapeutic effect of A is 39/110 and that of B is 61/110. This led to the opposite conclusion that B was superior to A.
This phenomenon in which the tendency of each group and the tendency of combining them are reversed is called Simpson's ( ).