I have heard some epidemiologists say that you must choose a random effects or fixed effects model and utilize for every outcome in your paper. Is this widely accepted or can you choose the model that fits best for a particular outcome in the paper?

More Greg Marchand's questions See All
Similar questions and discussions