On Sun, 25 Feb 2001, Richard Geller wrote:

> I think that there is more bias in:
>
> a) whether a study gets funded in the first place, and
>
> b) whether a study gets started and then killed by the researchers (in which
> case you never hear about it)
>
> than in the study, once it gets completed and published. My sense is that
> there is more bias hidden in a) and b).

I agree completely.  This is not to say that junk studies don't
make it to press.  Obviously they do.  But to dismiss the results
of a study because of who funded it is cultic thinking.  To be
suspicious and to look for the way in which biases might be built
into the research design, on the other hand, is a very good idea.

To go back to the example of Stefansson's Bellevue experiment in
1928.  That study *was* funded by the American Meat Packers
Institute.  The results were generally favorable, but not
entirely so.  Stefansson and Andersen were in negative calcium
balance for the entire year, and that was not suppressed from the
published reports.  Andersen's cholesterol did become extremely
elevated, which even then was recognized as a concern.  Ray
Audette may be right that this was caused by his coffee
consumption, but unless we also know that he *stopped* drinking
unfiltered coffee at the end of the experiment, we can't be sure.

Todd Moody
[log in to unmask]