I am working with longitudinal data where the outcome is the number of steps per minute. My LMM fit would look like:

lme(step ~ predictor, random = ~1|person, data = df, na.action = na.omit, method = "ML")

Another component of my project is power simulation from simulated data, and I am confused with how I can compute the variance and covariances for step data that are needed to compute "Sigma" from the mvrnorm function.

My sigma function current looks like

Sigma <- matrix(c(varSteps,rep(covSteps,(nRep))),nRep,nRep,byrow=T) 

So I am trying to compute "varSteps" and "covSteps" that are needed to construct a sigma matrix that will be used in my mvrnorm function.

How can I compute these two using R or SAS? Your help will be greatly appreciated.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.