topic Re: Logistic Regression With Multiple independent variables in SAS Enterprise Guide

Logistic Regression With Multiple independent variables

samrae — Wed, 26 Mar 2014 14:37:01 GMT

I am trying to do a logistic regression in sas enterprise guide 5.1. I have 1 dependent variable and 13 independent variables. out of the 13 independent variables 12 are possible confounders and I am specifically looking at one variable but I still want to get the crude odds ratio and 95% ci for all independent variables.

Re: Logistic Regression With Multiple independent variables

PaigeMiller — Wed, 26 Mar 2014 15:42:11 GMT

What is the specific question you want answered?

Re: Logistic Regression With Multiple independent variables

samrae — Wed, 26 Mar 2014 15:57:05 GMT

I want to know how to input it correctly. For some reason I list my dependent variable and all 13 variables under the explanatory variables section including the main independent variable. One of my variables is Sex which is not a numeric variable so i placed it under the "group by analysis" tab however i need it to generate an odds ratio and 95% ci as well in relation to my dependent variable

Re: Logistic Regression With Multiple independent variables

samrae — Wed, 26 Mar 2014 15:58:33 GMT

After doing all of that Ii get an error message saying there were no observations or something along those lines.

Re: Logistic Regression With Multiple independent variables

PaigeMiller — Wed, 26 Mar 2014 16:08:31 GMT

Exact details would help.

The exact error message, word-for-word, verbatim, leaving nothing out, would be most helpful.

The full set-up that you provided EG would also be necessary.

Re: Logistic Regression With Multiple independent variables

samrae — Wed, 26 Mar 2014 16:50:12 GMT

I imported my data which was an excel file. I then sorted it. Then I clicked Task-> Regression-> logistics->I then put my dependent variable under the tab designated for it. Then the rest of my variables I placed under quantitative variables(this includes both confounders and my main independent variable). I then placed "sex" the only variable I have that isn't a numeric one under the section that says classification variables. then I went to the left window pane, clicked effects and placed my main independent variable in the effects column. then I ran it and the error message says

ERROR: EVENT=. category is not a valid response category.