Open jiangaimin1995 opened 9 months ago
design<-svydesign(data=data1115, id=~id, strata=~strata, weights=~weightage, nest=TRUE) svy.fit<-svyglm(outcome ~ age + gender + drugs + educations + income, family=quasibinomial(link="logit"), design = design) y<-predict(svy.fit) y<-as.numeric(y)
y<-exp(y)/(1+exp(y))
And then, how should I get the OR and 95%CI?
design<-svydesign(data=data1115, id=~id, strata=~strata, weights=~weightage, nest=TRUE) svy.fit<-svyglm(outcome ~ age + gender + drugs + educations + income, family=quasibinomial(link="logit"), design = design) y<-predict(svy.fit) y<-as.numeric(y)
transform into probability
y<-exp(y)/(1+exp(y))
And then, how should I get the OR and 95%CI?