Most appropriate tests for variables?

#1
Hi all,

I'm hoping I can get some guidance around test selection for variables. My dependent variable is categorical (yes/no), and I have three independent categorical variables, two are ordinal (age and a grading) and one is nominal (ethnicity). Is there a test that is most appropriate for these? I've done some initial reading and have seen ordered logistic regression or factorial logistic regression as options for 2 or more categorical variables depending on the type of categorical variable, however am I on the right track with this? Or are there other options? I'm wanting to see if the level of grading increases the risk of disease diagnosis, and if ethnicity and age also increase the risk.

My stats is quite rusty so I'm having to revise, so any advice would be fantastic.
 

obh

Active Member
#2
Hi Luke,

I assume your question is what regression model to use?

When you choose the regression model you should look on the dependent variable (Y).
In your case it is a binary dependent variable, yes/no, so you should choose the logistic regression.