When regressing a dependent variable y on some feature vector x with a standard linear regression, is there any correction in place for multiplicity or is this just not relevant in this case?
The motivation for this question is that when I run several regression specifications (e.g. with or without interaction effects), I know that I have to perform some correction for the multiple testing to correct confidence intervals and p values (e.g. Benjamini-Hochberg Correction).
Now when I have a feature vector with e.g. 10 different independent variables, is estimating their coefficients equivalent to multiple testing?
