How can I provide unique cluster IDs for analysis with MPlus?
The SESTRAT within SECLUSTR values are currently set up as strata with a value of either 1 or 2 in the cluster or SECU variable. They are unique records within those variables but Mplus requires a unique cluster value within the strata. It is easy to just renumber and create new variables.
For example, in a data set with 100 records and SESTRAT ranges from 1-20 (just an example) and there are values of 1 or 2 in SECLUSTR then they could be renumbered to NEWSTRAT in 1-20 and NEWCLUSTR in 1 to 100 the total number in each of the NEWSTRAT variables.
Old Strata, Old Cluster, New Strata, New Cluster
1 1 1 1
1 1 1 2
1 2 1 3
1 2 1 4
2 1 2 5
2 2 2 6
20 2 20 100
Mplus requires unique strata/cluster combinations so this method allows you to define the records within Strata/PSU without changing the basic complex sample variable structure or meaning. The user could use the define statement in Mplus or do this outside of Mplus in the software of choice.