This is a linkpost for https://andershuitfeldt.net/2017/01/25/odds-ratios-and-conditional-risk-ratios/
I wrote this on my personal blog; I was reluctant to post this to Less Wrong since it is not obviously relevant to the core interests of LW users. However, I concluded that some of you may find it interesting as an example of how the academic publishing system is broken. It is relevant to Eliezer's recent Facebook comments about building an intellectual edifice.
I thought it was interesting -- and frustrating for you. I haven't invested the time into proving to myself you're right, but in the case that you are I hope you're able to get someone to verify and lend you their credibility.
Why do you think two senior biostats guys would disagree with you if it was obviously wrong? I have worked with enough academics to know that they are far far from infallible, but curious on your analysis of this question.
Why do you think two senior biostats guys would disagree with you if it was obviously wrong? I have worked with enough academics to know that they are far far from infallible, but curious on your analysis of this question.
Good question. I think a lot of this is due to a cultural difference between those of us who have been trained in the modern counterfactual causal framework, and an old generation of methodologists who felt the old framework worked well enough for them and never bothered to learn about counterfactuals.
Update: The editors of the Journal of Clinical Epidemiology have now rejected my second letter to the editor, and thus helped prove Eliezer's point about four layers of conversation.
I wrote this on my personal blog; I was reluctant to post this to Less Wrong since it is not obviously relevant to the core interests of LW users. However, I concluded that some of you may find it interesting as an example of how the academic publishing system is broken. It is relevant to Eliezer's recent Facebook comments about building an intellectual edifice.