I am using Stata 14/IC - I have successfully conducted a multiple imputation using chained equations. I am now trying to run xtmixed models in three stages
an empty model, a model with only individual predictors and a model with individual and school level predictors.
This is my script -
** MODEL 1 – empty model
mi estimate: xtmixed tku_a || ïs_num:
estimate store em1
estat icc
** MODEL 2 – ind level only
mi estimate: xtmixed tku_a gender_num yip || ïs_num:
estimate store em2
estat icc
**lrtest
lrtest em1 em2
** MODEL 3 – school level (all levels)
mi estimate: xtmixed tku_a gender_num yip scty dmhl_rc tsel tats t_total b_sum || ïs_num:
estimate store em3
estat icc
**lrtest
lrtest em1 em3
lrtest em2 em3
I have two problems - the lrtest does not appear to be working and I understand the estat for ICC is not possible using imputed data.
Any suggestions on these two issues?
Related Posts with using lrtest to compare an empty, individual level and two-level xtmixed model after multiple imputation
Reducing repeated responses to find actual sample size from a multiple imputed datasetHi there I am hoping someone can advise me on this complex dataset that is derived from a dual frame…
Estimate GARCH-DCC with asymmetriesHas anyone had experience estimating GARCH-DCC models with asymmetries (GJR for example)? Can this b…
bysort queryHello, I am trying to use bys to generate sequential and total number of Lines of treatments by pati…
Multiple time-failure analysis, event-specific coeffcients?Dear all, I am using stcox along with a stratification variable to estimate the hazard functions fo…
How to create control group against fraud firms?Dear all, I want to create a control sample of non-fraud firms against firms that committed fraud. I…
Subscribe to:
Post Comments (Atom)
0 Response to using lrtest to compare an empty, individual level and two-level xtmixed model after multiple imputation
Post a Comment