Response to Invited Commentary Pischon et al. Respond to ‘‘Variable Selection versus Shrinkage in Control of Confounders’’

We read with great interest Dr. Greenland’s invited com-mentary (1) about variable selection to control for confound-ing in observational studies. We agree with Dr. Greenland that the identification of confounders should be based primar-ily on background knowledge and not on significance testing. However, our proposed method (2) is not meant primarily as a variable selection procedure. Currently, relative risk esti-mates are commonly presented from nested models with in-creasing complexity of covariate use (3). This is not caused predominantly by the uncertainty of selecting the proper model, but rather by the interest to quantify the relative effect of adjustment for specific covariates on risk estimates. For example, relative risks from a multivariate model for a spe-cific nutrient might be compared with those from a model with additional adjustment for other nutrients or foods t