2
$\begingroup$

We are fitting a two-level logistic regression model with continuous variables in Level 1 and no variables in Level 2, except the groups.

We considered only a random intercept in the model, but it is possible to consider a random slope for one or more independent variables.

What is the best method to center the variables in Level 1, by the total mean or by the group means?

Improve this question
$\endgroup$
4
  • $\begingroup$ Not sure if this is new to you or not, but this paper is great on the topic osf.io/mt5z3/download $\endgroup$
    – Phil Assheton
    Commented May 28, 2021 at 21:22
  • 1
    $\begingroup$ There is quite a large literacture on this. What is the purpose of your model ? What is/are your research question/s ? $\endgroup$
    – Robert Long
    Commented May 29, 2021 at 10:48
  • $\begingroup$ Please can you provide more details about this ? $\endgroup$
    – Robert Long
    Commented Jun 5, 2021 at 18:12
  • 2
    $\begingroup$ I’m voting to close this question because the OP has not answered the questions in the comments and has not been on the site in years. $\endgroup$
    – Peter Flom
    Commented Dec 17, 2023 at 12:21

0

Reset to default