Page 1 of 1
duplicate cases
Posted: Tue May 14, 2013 5:26 pm
by Katerinaki
Hi there,
I am working on TIMMS data, and I saw that some countries have the same code in IDTEALIN variable. MlWin will think that there are different classes (1 level) (with same label) in different countries(2 level) or same classes (with same label) in different countries.
Hope make sense my question
Thanks a lot...
Katerina
Re: duplicate cases
Posted: Tue May 14, 2013 9:21 pm
by ChrisCharlton
I am not familiar with the TIMMS dataset however, except for cross-classified MCMC models, MLwiN uses changes in identifiers rather than their values to determine the unit numbers. For example for the following data:
Code: Select all
student school
1 1
2 1
1 2
2 2
1 1
2 1
1 2
2 2
MLwiN will interpret this as there being 4 schools, with 2 students in each.
This will also be the case for the following patterns.
Code: Select all
student school
1 1
2 1
3 2
4 2
5 1
6 1
7 2
8 2
Code: Select all
student school
1 1
2 1
1 2
2 2
1 3
2 3
1 4
2 4
Code: Select all
student school
1 1
2 1
3 2
4 2
5 3
6 3
7 4
8 4
Having said this it is still probably worth creating unique identifiers in case you wanted to run a cross-classified model or you took a sub sample that caused the identical identifiers to end up next to each other.
You can also check that the data hierarchy and number of units at each level are what you are expecting through the Model->Hierarchy Viewer window.
For further information on data ordering see
http://www.bristol.ac.uk/cmm/software/s ... html#order.
Re: duplicate cases
Posted: Wed May 15, 2013 8:55 am
by Katerinaki
Hi Chris
Thanks for your reply, I will follow your advice and I will create unique identifiers.
Katerina