Propensity Scores: Observational Data and Wishful Thinking
/Propensity score methods attempt to make observational data look like a randomized trial. But there are some big limitations that need to be considered before we jump on the bandwagon. For the video version, click here.
Read MoreAdjustment is a lie. Or how I've never seen a linear relationship in any aspect of biology.
/We researchers "adjust" our observational data all the time. And we should But we often apply a simple linear model when we adjust, which is almost always wrong. Read more:
The Methods Man: Devil in the (Adjusted) Data | Medpage Today.