In article , Bill Oliver wrote:
In article , Bill Oliver wrote:

The paradigm for glyphosate is the Non-Hodgkins lymphoma example, which
is also association with the development of Non-Hodgkins lymphoma in
which univariate analysis showed a weak association between glyphosate

should read

The paradigm for glyphosate is the Non-Hodgkins lymphoma example, in
[delete line]
which univariate analysis showed a weak association between glyphosate

Cut and paste error




My bad. I was commenting on the *other* article you posted the abstract
for -- the Minnesota one, not the Ontario one. So here's the comment
on the Ontario one. Since I went on at length about the Minnesota
one, I'll just let the authors provide the caveats for the the
Ontario one:


"Although this study is one of the first to collect and analyze
detailed information on the timing and types of pesticides used on
farms and reproductive outcomes, several limitations suggest that our
findings be interpreted with caution. Because dose information was not
available, misclassification of exposure is likely. Many factors
including the pesticide formulation, application conditions, handling
practices, and interindividual differences in absorption,
distribution, metabolism, and excretion of the products or metabolites
will lead to variability in the degree of exposure. Because the
farmers used many different pesticides during the study and our sample
size was limited, findings may be unreliable, particularly for
multiple pesticide interactions. Because pesticide products were
reported primarily by the farm applicator or husband, differential
recall of pesticide exposure by the mother is not likely to be a
problem in this study; however, some nondifferential recall of
pesticides and spontaneous abortions is likely. Because the analyses
were designed to generate, not to test, hypotheses, and multiple
comparisons were conducted, results should be interpreted with care
and tested in other studies."

So, aside from the problem that the kind of exposure might be
wrong, the dosage might be wrong, the interactions might be
unknown, and the actual association between the use of
the pesticide and the abortion might be wrong, and that
the study doesn't actually test what you claim it proves,
it's rock solid.

Sometimes it does pay to read the articles, eh?

billo