Search found 49 matches

by Harvey Goldstein
Mon Mar 03, 2014 11:57 am
Forum: Realcom user forum
Topic: 2 level vs. single level imputation model
Replies: 6
Views: 6805

Re: 2 level vs. single level imputation model

If your model of interest is 2 level and you have missing values in the response then when imputing these you should run a 2 level imputation model.
by Harvey Goldstein
Mon Nov 11, 2013 5:39 pm
Forum: Realcom user forum
Topic: Missing values in the imputation model
Replies: 2
Views: 4167

Re: Missing values in the imputation model

This is what realcom impute is designed to do. Please read the manual for examples and further details.
Do come back if you get stuck
Harvey Goldstein
by Harvey Goldstein
Fri Sep 27, 2013 10:08 am
Forum: Realcom user forum
Topic: Imputation very slow
Replies: 7
Views: 7940

Re: Imputation very slow

Yes, REALCOM is slow. I don't see ant shortcut to going through the whole fitting process. All will be speeded up when STATJR starts operating in November. Re the response variable with missing values - I would put it in as a response in REALCOM.
Harvey Goldstein
by Harvey Goldstein
Thu Jun 13, 2013 6:25 pm
Forum: Realcom user forum
Topic: Conceptual question
Replies: 1
Views: 3042

Re: Conceptual question

That's correct you will need to re-impute
Harvey Goldstein
by Harvey Goldstein
Thu Jun 13, 2013 6:24 pm
Forum: Realcom user forum
Topic: Limiting imputation values to the range of the variable
Replies: 2
Views: 4492

Re: Limiting imputation values to the range of the variable

At the moment unfortunately this cannot be done in REALCOM. We do have very experimental software that will do this but not yet suitable for release. We hope to make this facility available later this year and will announce it when ready.
Harvey Goldstein
by Harvey Goldstein
Wed Apr 24, 2013 1:26 pm
Forum: Realcom user forum
Topic: non-linear relationship & restricting imputation of variable
Replies: 1
Views: 3400

Re: non-linear relationship & restricting imputation of vari

Regarding your first Q I would suggest simply using the squared term for each variable. Since this is just for imputation purposes and assuming that the relationship is very close to linear for non-BMI it should be fine. Re second question option a) would be fine but make sure that you treat number ...
by Harvey Goldstein
Mon Apr 08, 2013 3:55 pm
Forum: MLwiN user forum
Topic: TIMSS - plausible values
Replies: 2
Views: 5580

Re: TIMSS - plausible values

The PVs are essentially multiple imputations. These should be combined using what ar eknown as Rubins rules and there should be guidance from TIMSS on this. You cannot treat them in any sense as a 2 level structure. If you want to do multilevel multiple imputation using individual records you can tr...
by Harvey Goldstein
Mon Apr 08, 2013 3:50 pm
Forum: Realcom user forum
Topic: Data structure in Structure equation model
Replies: 1
Views: 3990

Re: Data structure in Structure equation model

For categorical predictors use dummy variable codes. Re you results this seems to be a data problem - I suggest you look for rogue values or linear dependencies among predictors.
Harvey Goldstein
by Harvey Goldstein
Thu Feb 28, 2013 3:00 pm
Forum: Realcom user forum
Topic: multiple questions
Replies: 4
Views: 5041

Re: multiple questions

Shane A number of interesting questions! You could include all the level 2 differently defined variables in 1 imputation run. However, MLwin is only currently set up to import the imputed datasets back into the model originally set up in mlwin so this isn't going to work very neatly - I suggest you ...
by Harvey Goldstein
Tue Feb 05, 2013 10:33 pm
Forum: Realcom user forum
Topic: Problem when imputing second level variable
Replies: 3
Views: 4955

Re: Problem when imputing second level variable

There is no way in the REALCOM version on the web site that you can set impute limits. I was not aware that there was a bug of this sort - certainly not in latest version. It is possible that there is a problem with STATA. So - check you have latest version. If still a problem I suggest that you pre...