Logistic Regression with Fixed effect

[mjalvarez]

Hi community, 

I have a question about using PROC GLIMMIX fixed effects for a logistic regression. 

I am trying to estimate the effect of race (0 v 1), region (1 v 2), and 4 other covariates (all variables are categorical) on the likelihood of being selected as chief monitor. I want to treat region as a fixed effect. 

This is my code - 

proc glimmix;
class race region;
model cr = race region sex year citizen schooltype newaoa
/ dist=binary link=logit ddfm=bw solution;
random intercept / subject=race;
run;

My primary question is, how can I print the odds ratio of each predictor?

As a secondary question, Is this procedure the best approach for my question? 

Thank you all! 

[PaigeMiller]

---

@mjalvarez wrote:

My primary question is, how can I print the odds ratio of each predictor?

---

You use the ODDSRATIO option of the MODEL statement.

As a secondary question, Is this procedure the best approach for my question? 

Seems like it should work, but you say all variables are categorical, yet only two variables are in the CLASS statement, this seems questionable. Do you have specific concerns about it?

[mjalvarez]

Hi Paige,
Thanks for the response! I added the ODDSRATIO option of the MODEL statement like you suggested. I realize I didn't add the covariates to the class statement which led to some odd odds ratios but I fixed it and all seems good. Thank you!