topic Odds Ratios from ordinal LR using surveylogistic in Statistical Procedures

Odds Ratios from ordinal LR using surveylogistic

RDD — Sat, 18 Feb 2017 19:16:36 GMT

I am running an ordinal logisic regression analysis where the outcome/idependent variable (Q169_2re) has three levels (1=little exposure, 2=moderate, 3=extreme). I am using the following code and I am unable to get odds ratio estmates for each level. I tried a contrast statement but it didn't work (I thought those were only for levels of depenent variables, but I could be wrong). Can someone point me in the right direction?

proc SURVEYLOGISTIC data=HRBM order=formatted;

weight weight1;

class Q169_2re (ref="1 little exposure") risk(ref="no risk")/PARAM=ordinal;

model Q169_2re=risk/CLODDS;

run;

Re: Odds Ratios from ordinal LR using surveylogistic

StatDave — Fri, 24 Feb 2017 19:30:40 GMT

I assume that you specified PARAM=ORDINAL because your response is ordinal. You never need to specify the response variable in the CLASS statement, and I recommend that you don't. Because you have PARAM=ORDINAL specified as a global option in the CLASS statement, it applies not only to the response but also your predictor. That's why you don't get odds ratios by default (though you could probably add an ODDSRATIO statement to get it). So, the following would be better code. Note that if you need to model the probabilities of higher values of the response, use the DESCENDING response variable option as shown.

proc SURVEYLOGISTIC data=HRBM order=formatted;

weight weight1;

class risk(ref="no risk");

model Q169_2re(descending)=risk/CLODDS;

run;

Re: Odds Ratios from ordinal LR using surveylogistic

RDD — Fri, 24 Feb 2017 20:06:38 GMT

Thank you