mean centering & variables scaling

[HyunJoo LEE]

hello. I am planning to use a mixed model when writing my paper. I have three questions:

1. When controlling covariates such as age, years of education, and gender, it is said that gender is treated as a dummy variable because it is nominal. What about age or years of education? Can I check centered?

2. When the level 2 independent variable is a group, mean centering cannot be performed, so can it be treated as a dummy variable in this case as well?

3. When I want to add covariates, is it correct to put the nominal variables in factors and the rest in covariates?

thank you for your help
have a good day.

[mcfanda@gmail.com]

1. sure, that is the default. With mixed model is always a good idea to center the covariates (although not necessary)
2. any factor is transformed in contrast variables and their coding is mean-centered (if do not select "dummy" as a coding system)
3. yes