multicollinearity check for logistic regression including binary and categorical variables

[Christian Frewer]

Hey everyone,

i have to analyse panel data for my thesis. I use Stata version 15.1.
I chose the logistic regression model for my empirical analysis as my dependent variable is a binary dummy variable.
Besides the main explanatory variable of interest, I added several other variables, some of them also dummy variables and also one categorical variable.


My current regression command: xtlogit dummy_1 IV_1 IV_2 ..... ibn.id_model .... IVn..., noconstant

dummy_1 is my dependent variable, IV1 my focus independent variable, IVn all other independent variables (including dummies), ibn.id_model is my categorical variable.


Question: how can i check for (multi)collinearity, especially regarding the categorical variable?

I read a few other posts about the -vif- command, which apparently cannot be used in combination with the option - noconstant -. Stata also doesn't let me use the -collin- command as "factor-variable and time-series operators (are) not allowed".

Thanks for your help in advance.

Kind regards,
Christian

[Marcos Almeida]

I wonder whether you have read this thread. If not, maybe it will be helpful.