Search found 49 matches

by Harvey Goldstein
Tue May 29, 2012 10:10 am
Forum: MLwiN user forum
Topic: Multiple imputation for longitudinal data
Replies: 2
Views: 4666

Re: Multiple imputation for longitudinal data

You can ceratinly use realcom for this. The data should be presented as a 2-level structure, i.e. sorted by measurement occasion within individual. Refer to manual for details. :roll:
by Harvey Goldstein
Tue Mar 06, 2012 1:46 pm
Forum: Realcom user forum
Topic: Convergence probelm - ???Error using ==> plus
Replies: 1
Views: 4069

Re: Convergence probelm - ???Error using ==> plus

Thanks for Picking this up Jamie. The bug has now been corrected and a new version is downloadable.
Harvey Goldstein
by Harvey Goldstein
Mon Feb 13, 2012 4:57 pm
Forum: Realcom user forum
Topic: Deviance to compare models
Replies: 2
Views: 4721

Re: Deviance to compare models

Realcom itself does not compute a 'fit' statistic such as DIC. If you are using realcom to do multiple imputation via max likelihood fits then you could consider using the average AIC or BIC fit statistics based on the usual deviance - but check the literature since these will not have the standard ...
by Harvey Goldstein
Thu Feb 02, 2012 4:42 pm
Forum: Realcom user forum
Topic: Problem specifying weights
Replies: 5
Views: 7180

Re: Problem specifying weights

The weighting facility no longer operates. We may reinstate this at some stage after further research currently being done.
Harvey Goldstein
by Harvey Goldstein
Tue Oct 18, 2011 10:10 am
Forum: Realcom user forum
Topic: imputation of ordered categorical level-2 variables
Replies: 2
Views: 4702

Re: imputation of ordered categorical level-2 variables

You seem to have uncovered a bug. I cant be sure but will have a look at this and see if I can see what is happening
Harvey Goldstein
by Harvey Goldstein
Sun Oct 16, 2011 2:46 pm
Forum: Realcom user forum
Topic: Error message for categorical variable imputation
Replies: 8
Views: 9191

Re: Error message for categorical variable imputation

If you send me the data you are trying to use with realcom I will look into this

h.goldstein@bristol.ac.uk
by Harvey Goldstein
Fri Oct 14, 2011 6:49 pm
Forum: Realcom user forum
Topic: Error message for categorical variable imputation
Replies: 8
Views: 9191

Re: Error message for categorical variable imputation

Regarding dummy variables, you must enter a multicategory response as a single variable with categories coded as integers starting at zero.
Harvey Goldstein
by Harvey Goldstein
Fri Oct 14, 2011 6:46 pm
Forum: Realcom user forum
Topic: How to make changes in MCMC estimation settings?
Replies: 3
Views: 4995

Re: How to make changes in MCMC estimation settings?

One of the issues with REALCOM is that it is slow - we are planning to do something about this with new software in the pipeline but this will not be until well into 2012. Reducing teh number of iterations will certainly shorten the time taken but with only a short burn in you may have poor mcmc cha...
by Harvey Goldstein
Tue Jun 28, 2011 5:28 pm
Forum: Realcom user forum
Topic: 'Cases in Use' After Imputation Query
Replies: 1
Views: 3404

Re: 'Cases in Use' After Imputation Query

Dear Ryan
Seems strange. If you could send the relevant data files (original data records i.e. mlwin worksheet indicating the variables used + some of your imputed datsets we can have a look.
best
Harvey Goldstein (h.goldstein@bristol.ac.uk)