On Fri, 31 Jul 1998, Richard Geller wrote:

> Todd, thanks for posting this abstract.
>
> Trouble is that this type of prospective study is notorious. This is
> exactly what made the Framingham study so flawed, researcher bias,
> participants' bias, etc. etc.
>
> Note that it is the "covariates" that are so hard to control. Perhaps
> the folks who ate a lot of nuts (or who _said_ they did) were in general
> more careful about what they ate, ate lower carb, ate fewer transfatty
> acids, etc. Who knows? And perhaps the men who ate red meat ate higher
> carb, ate more moldy fruit, or some other variate that the researchers
> didn't even _think_ about measuring/asking about/controlling or couldn't
> have controlled for.

You're right, of course, but often this sort of study is the only
game in town.  At least this study had the virtue of having a
large number of subjects, so that accidental correlations are far
less likely than they would be in a smaller study.  Adventists as
a group are indoctrinated to be health-conscious, according to
their own ideas of what that means.  Alcohol consumption is
strongly discouraged, as is tobacco smoking and even coffee
drinking.  In the absence of any specific reason to think that
nut-eaters were even *more* careful than other SDAs, I see no
reason to introduce a hypothetical confounding variable.

Clearly, a single study never settles anything.  But I'm unable
to find the "smoking gun" study that indicts peanuts.

Todd Moody
[log in to unmask]