Post

Oh no! Oh no!!

Is the problem in assuming that including more covariates are always better? Or in asserting that observational evidence + theoretical considerations are often the best option? I'm too dull to tell.