#correlation without #causation via #ColliderBias
a spurious relationship between the dependent (Y) & independent (X) variable can occur, when each effect a 3rd variable (C, #collider), and this 3rd is included in the linear model lm(Y~X+C)
Regardless of COVID, it seems that causal inference methods are finally entering the mainsteam.
Use of #DAGs & awareness of #ColliderBias and the #Table2Fallacy are skyrocketting! Even a general medical journal (JAMA) has now produced primers on these issues
But we are still desparately short of advice and guidance on how best to use causal inference methods for applied research; we need more funding for meta-science and methods translation!
This JAMA Guide to Statistics and Methods describes collider bias, illustrates examples in directed acyclic graphs, and explains how it can threaten the internal validity of a study and the accurate estimation of causal relationships in randomized clinical trials and observational studies.
#correlation without #causation via #ColliderBias
a spurious relationship between the dependent (Y) & independent (X) variable can occur, when each effect a 3rd variable (C, #collider), and this 3rd is included in the linear model lm(Y~X+C)