I'm using an ANOVA with nested random effects to estimate the overall mean in a data set. I also want to estimate the SEM or confidence interval of that mean and wonder how to do it.
I've seen that some take the sqrt(MSE/n), but am uncertain what n should be in this case. The number of measurements, or perhaps the error df?
Thankfull for any suggestions!
Advertise on Talk Stats