
@danlewer, @tdbrothers and colleagues explain why risk factors for health outcomes shouldn't be studied by putting several independent variables into a regression model and identifying which are statistically significant in a new Methods Primer. bit.ly/3JC6G7u









