s********9 发帖数: 74 | 1 Anyone knows how to do multiple comparasion for contingency table in SAS? | g**r 发帖数: 425 | 2 The easiest way, though somewhat conservative, is just do a bonfferoni
correction: i.e., divide your alpha by the number of tests you carry out.
This should be true for ANY kind of test. The basic idea of mc is very
simple: because if you do too many test, you'll always find something
significant, just by chance. To avoid claim something does not exist, one
would control the type I error to punish too many tests.
Again, it is conservative and the punishment is heavy, but it is VERY easy.
【在 s********9 的大作中提到】![](/moin_static193/solenoid/img/up.png) : Anyone knows how to do multiple comparasion for contingency table in SAS?
|
|