|
![](/i/fill.gif) |
On Tue, 14 Jul 2009 10:40:12 -0700, Patrick Elliott <sel### [at] npgcable com>
wrote:
>
>Usually, it works more like this:
>
>1. Someone does such a study.
>2. Raw data, when "properly" analyzed shows "no" significant results, or
>involves such a small, and homogeneous group that it is basically
>meaningless (and usually contradicted by other studies).
Referances?
If you have not read Ben Goldacre's column from The Guardian, "bad science".
http://www.badscience.net/
>3. All the data that contradicts the premise is thrown out, and its
>deemed a success, because what was "kept" implies it worked.
>
There are three kinds of lies: lies, damned lies, and statistics.
>The joke being that
[large snip]
placebo effect?
It workes you know.
--
Regards
Stephen
Post a reply to this message
|
![](/i/fill.gif) |