Dear All,
I try to simulate correlated binary data for a clinical research project.
Unfortunately, I do not come to grips with bindata().
Consider
corr<-data.frame(ID=as.factor(rep(c(1:10), each=5)),
task=as.factor(rep(c(1:5),10)))
[this format might be more appropriate:]
corr2<-data.frame(ID=as.factor(rep(c(1:10),5)),
tablet=as.factor(rep(c(1:5),each=10)))
Now, I want to add one column 'outcome' for the binary response:
* within each 'task', the probabilty of success (1) is fixed, (say
rbinom(x,1,0.7))
* within each 'ID', the outcomes are correlated (say 0.8)
How can I generate this column 'outcome' with the given proporties?
Many thanks for hints or even code!
Regards,
Christian
--
http://portal.gmx.net/de/go/dsl02
To answer my own question: The package mvtBinaryEP was helpful! Christian Lerch schrieb:> Dear All, > > I try to simulate correlated binary data for a clinical research project. > Unfortunately, I do not come to grips with bindata(). > > Consider > corr<-data.frame(ID=as.factor(rep(c(1:10), each=5)), > task=as.factor(rep(c(1:5),10))) > > [this format might be more appropriate:] > corr2<-data.frame(ID=as.factor(rep(c(1:10),5)), > tablet=as.factor(rep(c(1:5),each=10))) > > Now, I want to add one column 'outcome' for the binary response: > * within each 'task', the probabilty of success (1) is fixed, (say > rbinom(x,1,0.7)) > * within each 'ID', the outcomes are correlated (say 0.8) > > How can I generate this column 'outcome' with the given proporties? > > Many thanks for hints or even code! > > Regards, > Christian