Interesting read! I am thinking that maybe the ideas from the latter ones could be used to fit relative risk models with logistic regression (by duplicating the cases as non-cases), which is related to the first post. The motivation for that is here: https
1,071 followers
604 followers
@AnnMemmott A fab friend who is going to support me when I start the Brum Ucert at the end of Sept. First advice "Carole, you don't have to read the whole book when you're doing your research" So I am reading the paper you linked, then go off on a tangetnt
873 followers
@statsepi OR & RR are both valid association measures, but RR is easier to interpret. Problems occur when OR is interpreted as RR in high incidence contexts. Curiously, this topic fascinated me some years ago: https://t.co/UHOvAI1rPZ
929 followers
Nifty method paper on RRs and ORs for various logistic methods using simple but well thought out hypothetical data. https://t.co/Mx0uavqFOz