Hello It is possible to set up an lm() model where none of the categories of the categorical independent variable need to be used as references, I mean use the total mean instead. Thanks /R