Finding Posterior Odds Ratios (Bayesian logistic regression)?

SGPS — Sun, 14 Oct 2012 21:09:24 GMT

Hi everyone,

I used the GENMOD procedure to obtain a Bayesian logistic regression model but am unsure how to obtain the adjusted odds ratios.

This is the code that I used:

data Normalprior;

input _type_ $ intercept X1 X2 X3;

datalines;

Var 1e6 1e6 1e6 1e6

Mean 0 100 50 0

;

proc genmod data=mydata descending;

model y=X1 X2 X3 / d=bin link=logit;

bayes seed=1 coeffprior=normal (input=Normalprior) nmc=50000 thin=5

summary=all diagnostics=all plots=all outpost=postdata;

run;

Is there a way to obtain the adjusted odds ratios in SAS or do I have to manually calculate it (if so how)?

Thanks in advance!

Re: Finding Posterior Odds Ratios (Bayesian logistic regression)?

SteveDenham — Mon, 15 Oct 2012 13:02:47 GMT

Bad news, good news. The bad news is that you will probably have to calculate them in a data step. The good news is that the estimates that you need to exponentiate are easily obtainable from an ODS output statement. Try adding

ods output postsummaries=postsummaries;

to your code. The parameter estimates are on the log odds scale so exponentiating would give the adjusted odds ratios.

Hope this helps.

Steve Denham

topic Re: Finding Posterior Odds Ratios (Bayesian logistic regression)? in Statistical Procedures

Finding Posterior Odds Ratios (Bayesian logistic regression)?

Re: Finding Posterior Odds Ratios (Bayesian logistic regression)?