Jon Heron posted on Wednesday, May 26, 2004 - 12:15 am
Due to the approaching deadline for a grant proposal I am attempting to run before I can walk so I apologise if this is a basic question that could be answered with a bit reading.
I have been running one of the models listed in Muthen and Muthen 2002 (How to use a Monte Carlo study..) for the LGM with no missing data.
I wish to alter the covariate so that the split is not 50:50 (eg 75:25, 95:5) but everything I have tried either increases the power or has negligible effect. It seems to me that the variance of the covariate is mirrored in the residual variance for the intercept but this line has not proven very fruitful.