Tongyun Li posted on Wednesday, October 02, 2013 - 1:11 pm
I'm trying to run a multiple group analysis with Mplus, and the groups correspond to data collected at different time points. My problem is that the data set for time 1(group1) has 30 replicate weights and the data set for time 2(group2) has 80 replicate weights, with all other variables the same. Is it possible to run multiple group analysis in Mplus without combining data sets into one single data set? Since I cannot only use 30 out of 80 replicate weights for group2, is there another way to circumvent this problem?
I think the best approach would be to use 2400 replicate weights where each replicate weight is obtained as weight for group 1 (30 possibilities) and a weight for group 2 (80 possibilities) for a total of 2400 possible replicate weights.
Stephan Daus posted on Saturday, September 09, 2017 - 4:25 am
I have a similar problem. I have a dataset with many countries, each country using a varying number of replicate weights (mostly 75, some as few as 59). Thus some countries have missing values on the last replicate weights. How does Mplus deal with missingness on these rep.weights? Will it impact the SEs?