Yo!
Fri, Dec-26-03, 19:11
don't let ppl fool you that stat sig. is not important. it is
the *single* best report since it contains all information in
it: basically a measure of the effect size relative to its
error.++Eg., in logit if the SE of the RR is too high then the
p will not be significant.++Ppl who say sig is useless are
wrong or at least have not been able to convince *anyone*
-------------- ÀÎÅÍ³Ý Ä«¸®½º¸¶ KORNET -------------
the *single* best report since it contains all information in
it: basically a measure of the effect size relative to its
error.++Eg., in logit if the SE of the RR is too high then the
p will not be significant.++Ppl who say sig is useless are
wrong or at least have not been able to convince *anyone*
-------------- ÀÎÅÍ³Ý Ä«¸®½º¸¶ KORNET -------------