Selecting cases in a cluster with a minimum cluster size

[Paul Ginns]

Hi everyone,



I have several data files in which students are nested within higher order
structures (degrees, schools, faculties). I would like to restrict analyses
to cases where there is a minimum number of students in the same cluster
(e.g. at least 20 students in the same school). Can someone please suggest
how this might be done? I suspect it can be done through the DATA ->
AGGREGATE command but don't enough about it to do it myself.



Thanks in advance,



Paul

[Maguin, Eugene]

Paul,

This should be straight-forward with Aggregate. Read over the syntax
reference documentation. This should work

Aggregate outfile=* mode=addvariables/break=cluster/size=nu.

Then select cases with Size gt whatever your criterion is.

Gene Maguin