a thing does not have to be 100% true or 100% meaningless.
you are too quick to say "meaningless" when "complicated" is the right word.
3 breads is not a lot for US, maybe for Japan. It is likely a risk factor because the gut microbes co-evolving with Japanese (largely wheat free) would not be there to help digest gluten. But, in US they are present, and unless gut microbes are destroyed, generally no problem.
also relative risk regressions yield the unique contribution of each predictor, so the point about
studying those things alone is moot.
this was not an experiment with people assigned to diets. this was find about
all the different diet things in each group, and see what things are higher/lower in the UC group. and then report RR of each when controlling for the others.
it gets technical, but your main concern is just not a problem, and the study is not meaningless, it just does not have as much import as may first sound.
Post Edited (DBwithUC) : 12/24/2012 3:08:24 PM (GMT-7)