alternating logistic regression in sas

[tosin7_2000]

Hi
I am trying to run an alternating logistic regression is SAS to examine the relationship between Disease X symptoms and sleep. The study asks patients questions

Q1 :- Do you have Disease X Symptoms (Y / N response)

Q2 :- Do you wake up at night (Y/N)

These 2 questions were asked simultaneously and they subjects were asked these 2 questions repeated over a period of time. My objective is to test if there is an association between sleep and disease X symptoms.

I think this should be a bivariate regression. Example of my data attached




I modeled a ALR using the following code below

proc genmod data=both2 descend;

class subjid quest day ;

model response= day / dist=bin link =logit ;

repeated subject=subjid * question / LOGOR= exch ;

run;



My questions are

1) is my data set up write to model the ALR correctly in SAS

2) What is the correct interpretation of the Alphas. I think since i am using the overall odds for the marginal model a large odds will indicate an association but it seems several literature interprets it differently and the example on the sas website for ALR does not say

Thanks

[hlsmith]

Alphas represent the common log odds ratios for the cluster comparisons listed in the Log Odds Ratio Parameter Information, I believe when controlling for other listed idependent variables.