One important thing the ASA statement does not comment on is the role of the p-value in keeping bad papers _out_ of the literature. Imperfect it may be, but if that simple threshold were not out there, there's be a heck of a lot more irreproducible studies in the literature.
@@chadiao791 I do agree. The problem of nowadays publication bias, or at least a part of it, is due to the fact that people and journals tried to apply a dichotomus standard to papers between Statistically Significant and not; its the same as saying "good and bad".
who is that man that speaks @1:04:03? and did he publish the different problems remark?
David Cox.
@@jingangmiao8367 May he rest in peace
@@jingangmiao8367 p
One important thing the ASA statement does not comment on is the role of the p-value in keeping bad papers _out_ of the literature. Imperfect it may be, but if that simple threshold were not out there, there's be a heck of a lot more irreproducible studies in the literature.
Agreed. But it filters to a certain extent. There are so many junk papers out there with irreproducible results but with p value less than 0.05.
slrellison it seemed like you favored publication bias using p-value as a screening tool
@@chadiao791 I do agree. The problem of nowadays publication bias, or at least a part of it, is due to the fact that people and journals tried to apply a dichotomus standard to papers between Statistically Significant and not; its the same as saying "good and bad".
no! that is precisely the problem! good done research should not have anything to do with pvalues!!!!
Can OUP republish those Fisher’s classic three books ? So we can extract his original ideas,