right test? and how to interpretate?

[Tetsunochin]

Hi everyone,

I am not an expert in statistic and moreover in the use of JMP, so I am requiring your help.

I would like to find predictive factors for the success of a treatment (3 category :failure, partial and total success). More precisely I would like to test for example, the age (numeric continuous), the sex (character nominal, 2 category) or the localization of the disease (character nominal, 5 category).

What I have done is analyze>Fit Y by X:
I choose the success of treatment for the Y and the "predictive factor" for X. Depending of the predictive factor (character or numeric), I have a mosaic plot or a logistic fit respectively.

Can some1 tell me if what I have done is good or no? Is there other test to perform?
Also, I need help for the interpretation of the results. Which numbers do I have to look and what do they mean?

Any help is very welcome!

Regards,

Tetsunochin Message was edited by: Tetsunochin

[Doc_Duke]

What you are asking about is a ordinal logistic regression. PROC LOGISTIC can do it. However, most of us who read this are not JMP users and can't help with the syntax.

[Tetsunochin]

thank you anyway...I keep looking...