Hello, I'm running the following multilevel model: Usevariables = X1 X2 X3 X4 Y sdmvstra sdmvpsu newrfips weights2 cluster; categorical are Y; stratification=sdmvstra; cluster= sdmvpsu newrfips; weight= weights2; wtscale = cluster; within= X1 X2; between X3 X4;
Analysis: Type = complex twolevel;
model: %within% Y on X1 X2; %between% Y on X3 X4; and I obtain different parameter estimates and SE, depending on how I sort my data-set. Is there a specific way I should sort my data-set when implementing a two-levels model with two levels clustering? Thank you very much!
You don't need to sort your data. The sorting must be changing the data if you get different results. Do a TYPE=TWOLEVEL BASIC on both data sets to make sure the descriptive statistics are the same for both data sets.