Most important asssumption in linear models ... and the second most

Load packages library(tidyverse) library(mosaic) We are following here the advise of Gelman and Hill (2007, p. 46-47). Validity Quite obviously, the right predictors must be included in the model in order to learn something from the model. The “right” predictors means: avoiding the wrong ones, and including the correct ones. Easier said than done, particularly with a look to the causal inference aspects. Let’s turn to the next most important …