How would you do this analysis?

[Katy Pearce]

Suggestions welcome!

 

I have a group of people with traits A, U, E, and S (all continuous variables) — all of which are good predictors of a bunch of outcomes:

 

– Do P only.

– Do H only.

– Do M only.

– Do H + M both.

– Don’t do any of these.

 

(all of these are binary yes/no categorical variables.)

 

I’m trying to create a model (decision tree sort of) of those traits relating to the outcomes.

 

Any thoughts?

[Adam Fominaya]

I think you are into log linear analyses. However, i am not trained or experienced with this technique so the best i can do is pass along a link from my favorite people at ncstate:

http://faculty.chass.ncsu.edu/garson/PA765/logit.htm

[Author]

Posts