I have data from longitudinal study about BMI and other variables (that could have some effect on BMI) in children in different time points, and I want to impute missing values using function amelia() from R package Amelia.
Each variable has at least one missing value.
Some variables change with time (like family income), whereas some are fixed (like sex).
In several cases sex is not known.
My question is following:
How can I incorporate constraints about sex into the amelia code?
It means I need to include somehow facts when e.g. sex is "male" in one time point it must be "male" in the rest of time points.
Does anyone know how to stick these facts using amelia code?