News for hackers. Stuff that matters.
Re:Acetaminophen use in pregnancy linked to child's behavior... by paulajohnson on Tuesday August 16, 2016 @03:52AM (#12296088)
fifteenforty buries one of the biggest potential issues in his 4th para, so I'm going to highlight it: its not clear if this correlation was found post-hoc or not. If you measure 10 factors and 10 outcomes then you have a potential 100 different correlations. If you assume that any correlation where p<0.05 is significant then you are going to get 5 "significant" correlations just by chance. The summary talks about how they didn't find correlations with acetaminophen before and after pregnancy, so that's 3 distinct factors just there.
Re:Acetaminophen use in pregnancy linked to child's behavior... by DHMO on Tuesday August 16, 2016 @08:13AM (#12296891)
> they didn't find correlations with acetaminophen before and after pregnancy, so that's 3 distinct factors just there

No, it's perfectly valid for a drug to only have affect on the growing fetus but where there were not studies proving harm to the child or adult, e.g. Thalidomide.