Macro for logistic regression of outcome with more than two categories

[Frankzzf]

I'm estimating odds ratio between m_pattern (three categories) and c_pattern (three categories) with the following code.

I found macro for logistics regression with ordinal outcomes, but I failed to find macro for outcomes with more than two categories.

I wonder if there are macro for batch analysis and auto OR table ouput?

proc logistic data=diet.dietary;
class
m_pattern2 (param=ref ref="1");
model c_pattern2 (ref="1")=m_pattern2 /link=glogit rl cl;
run;

[Ksharp]

I don't understand what you are trying to do .

You should know if your response/Y variable is nominal variable,you would get TWO model separatedly (1 vs 3 and 2 vs 3 if take 3 as a reference level).

And check the example of Doc:

Nominal Response Data: Generalized Logits Model

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/statug/statug_logistic_examples04.htm

[Ksharp]

And better post it at Stat Forum, since it is a question about Statisical Model/Theory.
The experts of Statistic @lvm @StatDave @SteveDenham would see it.

[PaigeMiller]

PROC LOGISTIC should do this without a macro. What is it you want that PROC LOGISTIC doesn't do?

[Frankzzf]

I apologize for not being clear.
i mean a marco that could output word documents with created OR tables from logistic procedure

[PaigeMiller · Accepted solution]

---

@Frankzzf wrote:
I apologize for not being clear.
i mean a marco that could output word documents with created OR tables from logistic procedure

---

Macros are not needed. You can use ODS RTF or ODS WORD. Example:

ods rtf file="myfile.rtf";
proc logistic data=...;
...
run;
ods rtf close;