The best of causal inference
The Effective Statistician - in association with PSI
00:00
What Are Your Key Take Away When Analyzing Observational Data?
Good study design is good a study design, and that it otlies both torandemise trials and to observation an studies. Anything that we do in a randomis trile should be done also in an observation stie though. Never compare carriageservice wis, never useas o use person time comparisons. In both cases, we need to adjust for compounders, both at base line and post base clain. But if we think about estimating the podual effect in a runmi trial, or the observational analogue of the othepeprotable effect in an observation study, they are exactly the same.
Transcript
Play full episode
