On 2 January 2011 21:50, Milos Rancic <millosh(a)gmail.com> wrote:
No really it isn't. Once you remove the waffle it really boils down to
"[[Publication bias]] exists" (while falling for it itself since it
fails to mention the clear publication bias in Funnel plots) and
"[[Regression toward the mean]] exists". Both of these are well known
and there are a range of statistical and methodological attacks to
deal with them.
The article then goes of the deep end and starts attacking some
strawman (some version of experimentalism by the look of it but that's
unclear) version of the scientific method.
--
geni