Edge is a forum where great minds can meet and discuss their ideas; the quality of the content is first rate, though unlike a TED talk, they are less structured, less tight, and thus, more fill.
In this video, Richard Nisbett talks about why he dislikes multiple regression analysis, which in fact is a dislike for the abuse of the tool; and for poor research design. Worth listening to, also when he goes on to talk about why the lack of replication need not be such a bad thing in the social sciences, as it helps point to the importance of contextual factors. It is the wealth and range of experiments that show the robustness of a concept, not a single example of it.
Worth watching 🙂
In the case of health data, the big problem is something that’s come to be called the healthy user bias, because the guy who’s taking Vitamin E is also doing everything else right.