Re:Google workers at Superfund site exposed
by carlob on Saturday March 30, 2013 @03:08PM
(#5466325)
As far as I understand only women that were older (≥ 38) had a significant effect from being exposed (on top of the increased risk due to being older).
You are misunderstanding the statistical theory. The xkcd comic was making a joke about misused priors. You have no reason to believe that this paper used a wrong prior and peer reviewers overlooked it, nor have you examined the choice of p-value.
So, they actually found no correlation with TCE exposure. There found a correlation with one of the subgroups [TCE + other variable], one out of many, that they selected after examining the data.
There's an XKCD about this:
http://xkcd.com/882/
Or did I misunderstand something?