3 ms·
It looks like they have found a stastically significant correlation with a very small effect size, using an extremely simple regression model adjusting for only
by zmmmmm 6y ago
It looks like they have found a stastically significant correlation with a very small effect size, using an extremely simple regression model adjusting for only a handful of covariates, looking across, presumably, tens of thousands or maybe even hundreds of thousands of SNPs.
It seems highly irresponsible to title such a study in a way implying a mechanistic or causal link. I expected to find a whole lot of discussion in the post about all the other evidence supporting it, but nope - they shoved their data into a stats package, and put the result straight into a blog post with a misleading title. They don't even discuss if they adjusted for multiple testing.
There's peer reviewed science, then there are pre-prints pending review. This is far far below either of those.
- op03 6y agoAs long as your performance in the Social Media Metrics driven arms race for Attention increasingly determines what funding/employees/clients you attract, this kind of behavior will continue and escalate.
- dekhn 6y agoMaybe the title for this post has been changed but when I read it, and read the PR, they are both very cautious to just note that this Evidence For, not Proof Of, or, mechanism or causality.