Date: Jun 20, 2013 10:09 AM
Author: Luis A. Afonso
Subject: Common sense and NHST
Common sense and NHST
Naive common sense sometimes doesn´t work
To deal smoothly with random events is an entire time-life endeavor, and it is, doubtless, the hardest part to perceive in order to be pretty comfortable, if so, with Data Processing, I mean.
Let be the ?counter sense?, but although true situation: It can happens that two no-significant p-values, issued from two independent NHST´s tests, can consist in a highly significant over all result.
Common sense says contrarily; if I found a no-significant value, and a second time too, common sense says that I got a strengthen evidence not to reject the null. NOT NECESSARILY, of course as we will show soon.
Luis A. Afonso