Hi there,

 

I dont know if you get any answer to your question but I just saw your question and will try to give some explanation.

 

Log-linear models are used to model expected frequencies or probabilities in contingency tables without distinguishing between independent and dependent variables. It is used for modelling categorical variables. In summary you model the counts in contingency tables and make inferences about the main effects or interaction effcets of categorical variables in contingency tables.

 

Logistic regression, on the contrary, is used for modeling dichotomuous categorical or ordered varaiables. In that case you have dependent variable which may be dichotomuous categorical variable or ordered variable. Independent variables may be both categorical or continuous.

 

These are the basic differences between two models. I hope those helps.

 

Best

Aylin

vrajak_msu <vrajak_msu@...> wrote: