Page 2 of 2

Re: How to replicate Stata do-file of MLRA using "runmlwin", please advise

Posted: Mon Mar 04, 2019 10:21 am
by ChrisCharlton
In your first set of models you have specified a level-2 variance, the second set is a 1-level model. In the second set of models you have chosen RIGLS instead of IGLS for the starting values and then set a seed, burnin and number of iterations for the MCMC chains. In the first set these will be set to the default values of seed=1, burnin=500 and iterations=5000 for the model chain.

If you want a picture of the equation to paste into Word you need to remove the nopause option from your command. When MLwiN opens and displays the equation that you want to copy then you need to use the menu Edit->Copy Screenshot (or the keyboard combination Control-G). This will put a copy of the equation on the clipboard as an image that you can then paste into other software such as Word.

Re: How to replicate Stata do-file of MLRA using "runmlwin", please advise

Posted: Wed Mar 10, 2021 3:26 pm
by ManuelDewez
Hello Hamza,

I am using runmlwin for a similar analysis based on the paper by Melro et al that you quote.
Did you manage to produce AUC and ROC curves?

Thank you.

Manuel