The Proof with Simon Hill cover image

Carnivore diet, veganism and nutrition misinformation | Dr Alan Flanagan

The Proof with Simon Hill

00:00

Randomization Is the Key to Causal Inference in a Randomized Controlled Trial

The assumption of randomization that makes people comfortable with causal inference is that any unknown covariates are assumed to be equally distributed between groups. A well-conducted cohort study conceptually can achieve the same thing if we know what those potential confounders are. This is a great point because on Twitter you often see people say, well, if it's an observational study that someone shared with a great multivariate analysis, really, the researchers were well aware of like confounding and adjusted for it. But this is really important because if we've got prior knowledge that there are things that influence an outcome, we use that prior knowledge well.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
 Get the app