I am trying to determine the needed sample size for a Logistic Mixed Effects Model powered at 90% with a type I, alpha, error of 5%. I don't have pilot data and there aren't similar studies that have repeated measures. I do have estimates for a logistic regression without repeated measures, including an Odds Ratio of 0.83 and a response probability of 6.9%. How can I estimate my sample size in SAS? Here is the code for the model I will be running: proc glimmix data=ecig method=quad oddsratio; class ID Year(ref='2018'); model vape(Event='1')=Year /s dist=binary link=logit ddfm=none chisq; random intercept/subject=ID; output out=glmxout predicted(blup ilink)=predprob; run;
... View more
