Ryan Marek posted on Wednesday, January 23, 2013 - 3:13 pm
We'd like to fit an HLM model to our data. We have three time points (time nested under person). We were told that we had to do last observation carried forward from our 1st time point to 2nd if our data was missing. Is this necessary? I thought as long as we have data on one time point, EM is used to estimate missingness.