Thomas Lumley • 2/11/2021

Co-linearity

This article critically examines multicollinearity in regression analysis, arguing that standard textbook diagnostics are often unhelpful. It explains the difference between coefficients in simple vs. multiple regression models and emphasizes that domain knowledge, not just diagnostics, is crucial for correct model specification and interpretation. The piece also covers implications for prediction tasks.

0 comments

#data analysis #statistics #Regression