cross-classified/cross-nested model: How to get all relevant random components

Welcome to the forum for runmlwin users. Feel free to post your question about runmlwin here. The Centre for Multilevel Modelling take no responsibility for the accuracy of these posts, we are unable to monitor them closely. Do go ahead and post your question and thank you in advance if you find the time to post any answers!

Go to runmlwin: Running MLwiN from within Stata >> http://www.bristol.ac.uk/cmm/software/runmlwin/
Post Reply
johannesmueller
Posts: 2
Joined: Sun Jul 29, 2018 12:37 pm

cross-classified/cross-nested model: How to get all relevant random components

Post by johannesmueller » Sun Jul 29, 2018 12:46 pm

Dear Madam or Sir,

I am running a cross-classified / cross-nested model in runmlwin and have some issues with the cross-classification for which I am seeking your help:

Code: Select all

runmlwin DV cons, level4(year: cons)  level3(Geo1: cons) level2(Geo2: cons) level1(ID: cons) mcmc(cc) initsb(b) nopause
Geo1 and Geo2 are cross-classified, i.e. it is NOT say counties within states, but any combination of them is possible. At the Geo1 / Geo2 level, there is a time dimension, moreover, there are two non-hierarchically nested geographical identifiers, and a large number of individuals nested in Geo1, Geo2. Individuals are nested within years, but it is a repeated cross-section, not panel of individuals.

The challenge is: While I have specified mcmc(cc) to get the cross-classification, and while I read in this forum that automatically all cross-classifications are made, I only get the random parts specified above, but not f.e. yearXGeo2, yearXGeo1, Geo1XGeo2, which however would also be needed to account for nesting as far as I know?
Should I generate them manually? If they are not part of the output, were they corrected for in the initial estimation?

I have no particular interest in knowing about Geo1Xyear, but since individuals are more similar in the same geo1-year, I would want to have a random effect there, as well as for all other combinations that are relevant given the data structure.

Thank you very much for your time!

Best

billb
Posts: 97
Joined: Fri May 21, 2010 1:21 pm

Re: cross-classified/cross-nested model: How to get all relevant random components

Post by billb » Mon Jul 30, 2018 8:50 am

Hi Johannes,
Basically in MLwiN (and therefore RunMLwiN) the classifications fitted are the ones you specify - in this case the 4 given each with a set of random effects. If you want to include 'interaction classifications' you need to create the appropriate variables and declare them as additional levels. Depending on the size of your data you might consider that year is in fact a fixed effect and include it as either a growth curve or dummy variables in the fixed part of your model. You could then add random slopes for year into your geographic levels.
Hope that helps,
Bill.

Post Reply